how to know if an essay is made by chatgpt

How-To Geek

How to tell if an article was written by chatgpt.

While no method is totally effective, you can train yourself to spot telltale markers of AI writing — for now.

Quick Links

How to tell if chatgpt wrote that article, can you use ai to detect ai-generated text, tools to check if an article was written by chatgpt, train your brain to catch ai, key takeaways.

You can tell a ChatGPT-written article by its simple, repetitive structure and its tendency to make logical and factual errors. Some tools are available for automatically detecting AI-generated text, but they are prone to false positives.

AI technology is changing what we see online and how we interact with the world. From a Midjourney photo of the Pope in a puffer coat to language learning models like ChatGPT, artificial intelligence is working its way into our lives.

The more sinister uses of AI tech, like a political disinformation campaign blasting out fake articles, mean we need to educate ourselves enough to spot the fakes. So how can you tell if an article is actually AI generated text?

Multiple methods and tools currently exist to help determine whether the article you're reading was written by a robot. Not all of them are 100% reliable, and they can deliver false positives, but they do offer a starting point.

One big marker of human-written text, at least for now, is randomness. While people will write using different styles and slang and often make typos, AI language models very rarely make those kinds of mistakes. According to MIT Technology Review , "human-written text is riddled with typos and is incredibly variable," while AI generated text models like ChatGPT are much better at creating typo-less text. Of course, a good copy editor will have the same effect, so you have to watch for more than just correct spelling.

Another indicator is punctuation patterns. Humans will use punctuation more randomly than an AI model might. AI generated text also usually contains more connector words like "the," "it," or "is" instead of larger more rarely used words because large language models operate by predicting what word will is most likely to come next, not coming up with something that would sound good the way a human might.

This is visible in ChatGPT's response to one of the stock questions on OpenAI's website. When asked, "Can you explain quantum computing in simple terms," you get sentences like: "What makes qubits special is that they can exist in multiple states at the same time, thanks to a property called superposition. It's like a qubit can be both a 0 and a 1 simultaneously. "

Short, simple connecting words are regularly used, the sentences are all a similar length, and paragraphs all follow a similar structure. The end result is writing that sounds and feels a bit robotic.

Large language models themselves can be trained to spot AI generated writing. Training the system on two sets of text --- one written by AI and the other written by people --- can theoretically teach the model to recognize and detect AI writing like ChatGPT.

Researchers are also working on watermarking methods to detect AI articles and text. Tom Goldstein, who teaches computer science at the University of Maryland, is working on a way to build watermarks into AI language models in the hope that it can help detect machine-generated writing even if it's good enough to mimic human randomness.

Invisible to the naked eye, the watermark would be detectable by an algorithm, which would indicate it as either human or AI generated depending on how often it adhered to or broke the watermarking rules. Unfortunately, this method hasn't tested so well on later models of ChatGPT.

You can find multiple copy-and-paste tools online to help you check whether an article is AI generated. Many of them use language models to scan the text, including ChatGPT-4 itself.

Undetectable AI , for example, markets itself as a tool to make your AI writing indistinguishable from a human's. Copy and paste the text into its window and the program checks it against results from other AI detection tools like GPTZero to assign it a likelihood score --- it basically checks whether eight other AI detectors would think your text was written by a robot.

Originality is another tool, geared toward large publishers and content producers. It claims to be more accurate than others on the market and uses ChatGPT-4 to help detect text written by AI. Other popular checking tools include:

Most of these tools give you a percentage value, like 96% human and 4% AI, to determine how likely it is that the text was written by a human. If the score is 40-50% AI or higher, it's likely the piece was AI-generated.

While developers are working to make these tools better at detecting AI generated text, none of them are totally accurate and can falsely flag human content as AI generated. There's also concern that since large language models like GPT-4 are improving so quickly, detection models are constantly playing catchup.

Related: Can ChatGPT Write Essays: Is Using AI to Write Essays a Good Idea?

In addition to using tools, you can train yourself to catch AI generated content. It takes practice, but over time you can get better at it.

Daphne Ippolito, a senior research scientist at Google's AI division Google Brain, made a game called Real Or Fake Text  (ROFT) that can help you separate human sentences from robotic ones by gradually training you to notice when a sentence doesn't quite look right.

One common marker of AI text, according to Ippolito, is nonsensical statements like "it takes two hours to make a cup of coffee." Ippolito's game is largely focused on helping people detect those kinds of errors. In fact, there have been multiple instances of an AI writing program stating inaccurate facts with total confidence --- you probably shouldn't ask it to do your math assignment , either, as it doesn't seem to handle numerical calculations very well.

Right now, these are the best detection methods we have to catch text written by an AI program. Language models are getting better at a speed that renders current detection methods outdated pretty quickly, however, leaving us in, as Melissa Heikkilä writes for MIT Technology Review, an arms race.

Related: How to Fact-Check ChatGPT With Bing AI Chat

• PRO Courses Guides New Tech Help Pro Expert Videos About wikiHow Pro Upgrade Sign In
• EDIT Edit this Article
• EXPLORE Tech Help Pro About Us Random Article Quizzes Request a New Article Community Dashboard This Or That Game Popular Categories Arts and Entertainment Artwork Books Movies Computers and Electronics Computers Phone Skills Technology Hacks Health Men's Health Mental Health Women's Health Relationships Dating Love Relationship Issues Hobbies and Crafts Crafts Drawing Games Education & Communication Communication Skills Personal Development Studying Personal Care and Style Fashion Hair Care Personal Hygiene Youth Personal Care School Stuff Dating All Categories Arts and Entertainment Finance and Business Home and Garden Relationship Quizzes Cars & Other Vehicles Food and Entertaining Personal Care and Style Sports and Fitness Computers and Electronics Health Pets and Animals Travel Education & Communication Hobbies and Crafts Philosophy and Religion Work World Family Life Holidays and Traditions Relationships Youth
• Browse Articles
• Learn Something New
• Quizzes Hot
• This Or That Game New
• Train Your Brain
• Explore More
• Support wikiHow
• About wikiHow
• Log in / Sign up
• Computers and Electronics
• Online Communications

Check if Something Was Written by ChatGPT: AI Detection Guide

Last Updated: December 9, 2023 Fact Checked

How AI Detection Tools Work

Using ai detection tools, signs of chatgpt use.

This article was written by Stan Kats and by wikiHow staff writer, Nicole Levine, MFA . Stan Kats is the COO and Chief Technologist for The STG IT Consulting Group in West Hollywood, California. Stan provides comprehensive technology & cybersecurity solutions to businesses through managed IT services, and for individuals through his consumer service business, Stan's Tech Garage. Stan has over 7 years of cybersecurity experience, holding senior positions in information security at General Motors, AIG, and Aramark over his career. Stan received a BA in International Relations from The University of Southern California. There are 11 references cited in this article, which can be found at the bottom of the page. This article has been fact-checked, ensuring the accuracy of any cited facts and confirming the authority of its sources. This article has been viewed 42,810 times.

With the rising popularity of ChatGPT, Bard, and other AI chatbots, it can be hard to tell whether a piece of writing was created by a human or AI. There are many AI detection tools available, but the truth is, many of these tools can produce both false-positive and false-negative results in essays, articles, cover letters, and other content. Fortunately, there are still reliable ways to tell whether a piece of writing was generated by ChatGPT or written by a human. This wikiHow article will cover the best AI detection tools for teachers, students, and other curious users, and provide helpful tricks for spotting AI-written content by sight.

Things You Should Know

• Tools like OpenAI's Text Classifier, GPTZero, and Copyleaks can check writing for ChatGPT, LLaMA, and other AI language model use.
• ChatGPT often produces writing that looks "perfect" on the surface but contains false information.
• Some signs that ChatGPT did the writing: A lack of descriptive language, words like "firstly" and "secondly," and sentences that look right but don't make sense.

• The detection tool compares a piece of writing to similar content, decides how predictable the text is, and labels the text as either human or AI-generated.
• These tools also look for other indicators, or "signatures" that are associated with AI-generated text, such as word choice and patterns. [1] X Research source

• If an AI detection tool reports that a piece of writing was mostly AI-generated, don't rely on that report alone. It's best to only use AI detection tools if you've already found other signs that the writing was written by ChatGPT. [3] X Research source
• Running a piece of writing through multiple AI detection tools can help you get an idea of how different tools work. It can also help you narrow down false-negatives and false-positives.

• If you're evaluating a piece of writing for potential AI use, try searching the web for a few facts from the text. Try to search for facts that are easy to verify—e.g., dates and specific events.

• For example, if you're evaluating a cover letter for AI use, you might tell ChatGPT, "Write me a cover letter for a junior developer position at Company X. Explain that I graduated from Rutgers with a Computer Science degree, love JavaScript and Ruby, and have been working as a barista for the past year."
• Because ChatGPT is conversational, you can continue providing more context. For example, "add something to the cover letter about not jumping right into the industry after college because of the pandemic."

Expert Q&A

• Cornell researchers determined that humans incorrectly found AI-generated news articles credible more than 60% of the time. [12] X Research source Thanks Helpful 0 Not Helpful 0
• If you're using a ChatGPT detection tool that identified writing as AI-written, consider that it may be a false positive before approaching the situation with the writer. Thanks Helpful 0 Not Helpful 0
• If you suspect ChatGPT wrote something but can't tell for sure, have a conversation with the writer. Don't accuse them of using ChatGPT—instead, ask them more questions about the writing or content to make their knowledge lines up with the content. You may also want to ask them about their writing process to see if they admit to using ChatGPT or other AI writing tools. Thanks Helpful 0 Not Helpful 0

You Might Also Like

• ↑ https://www.turnitin.com/blog/ai-writing-the-challenge-and-opportunity-in-front-of-education-now
• ↑ https://www.turnitin.com/blog/understanding-false-positives-within-our-ai-writing-detection-capabilities
• ↑ https://help.openai.com/en/collections/5929286-educator-faq
• ↑ https://www.npr.org/2023/01/09/1147549845/gptzero-ai-chatgpt-edward-tian-plagiarism
• ↑ https://app.gptzero.me/app/subscription-plans
• ↑ https://contentatscale.ai/ai-content-detector/
• ↑ https://copyleaks.com/api-pricing
• ↑ https://research.google/pubs/pub51844/
• ↑ https://help.openai.com/en/articles/6783457-what-is-chatgpt
• ↑ https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9939079/
• ↑ https://www.technologyreview.com/2022/12/19/1065596/how-to-spot-ai-generated-text/

About This Article

• Send fan mail to authors

Is this article up to date?

Featured Articles

Trending Articles

Watch Articles

• Terms of Use
• Privacy Policy
• Do Not Sell or Share My Info
• Not Selling Info

wikiHow Tech Help Pro:

Level up your tech skills and stay ahead of the curve

• Skip to main content
• Keyboard shortcuts for audio player

A college student created an app that can tell whether AI wrote an essay

Emma Bowman

GPTZero in action: The bot correctly detected AI-written text. The writing sample that was submitted? ChatGPT's attempt at "an essay on the ethics of AI plagiarism that could pass a ChatGPT detector tool." GPTZero.me/Screenshot by NPR hide caption

GPTZero in action: The bot correctly detected AI-written text. The writing sample that was submitted? ChatGPT's attempt at "an essay on the ethics of AI plagiarism that could pass a ChatGPT detector tool."

Teachers worried about students turning in essays written by a popular artificial intelligence chatbot now have a new tool of their own.

Edward Tian, a 22-year-old senior at Princeton University, has built an app to detect whether text is written by ChatGPT, the viral chatbot that's sparked fears over its potential for unethical uses in academia.

Edward Tian, a 22-year-old computer science student at Princeton, created an app that detects essays written by the impressive AI-powered language model known as ChatGPT. Edward Tian hide caption

Edward Tian, a 22-year-old computer science student at Princeton, created an app that detects essays written by the impressive AI-powered language model known as ChatGPT.

Tian, a computer science major who is minoring in journalism, spent part of his winter break creating GPTZero, which he said can "quickly and efficiently" decipher whether a human or ChatGPT authored an essay.

His motivation to create the bot was to fight what he sees as an increase in AI plagiarism. Since the release of ChatGPT in late November, there have been reports of students using the breakthrough language model to pass off AI-written assignments as their own.

"there's so much chatgpt hype going around. is this and that written by AI? we as humans deserve to know!" Tian wrote in a tweet introducing GPTZero.

Tian said many teachers have reached out to him after he released his bot online on Jan. 2, telling him about the positive results they've seen from testing it.

More than 30,000 people had tried out GPTZero within a week of its launch. It was so popular that the app crashed. Streamlit, the free platform that hosts GPTZero, has since stepped in to support Tian with more memory and resources to handle the web traffic.

How GPTZero works

To determine whether an excerpt is written by a bot, GPTZero uses two indicators: "perplexity" and "burstiness." Perplexity measures the complexity of text; if GPTZero is perplexed by the text, then it has a high complexity and it's more likely to be human-written. However, if the text is more familiar to the bot — because it's been trained on such data — then it will have low complexity and therefore is more likely to be AI-generated.

Separately, burstiness compares the variations of sentences. Humans tend to write with greater burstiness, for example, with some longer or complex sentences alongside shorter ones. AI sentences tend to be more uniform.

In a demonstration video, Tian compared the app's analysis of a story in The New Yorker and a LinkedIn post written by ChatGPT. It successfully distinguished writing by a human versus AI.

A new AI chatbot might do your homework for you. But it's still not an A+ student

Tian acknowledged that his bot isn't foolproof, as some users have reported when putting it to the test. He said he's still working to improve the model's accuracy.

But by designing an app that sheds some light on what separates human from AI, the tool helps work toward a core mission for Tian: bringing transparency to AI.

"For so long, AI has been a black box where we really don't know what's going on inside," he said. "And with GPTZero, I wanted to start pushing back and fighting against that."

The quest to curb AI plagiarism

Untangling Disinformation

Ai-generated fake faces have become a hallmark of online influence operations.

The college senior isn't alone in the race to rein in AI plagiarism and forgery. OpenAI, the developer of ChatGPT, has signaled a commitment to preventing AI plagiarism and other nefarious applications. Last month, Scott Aaronson, a researcher currently focusing on AI safety at OpenAI, revealed that the company has been working on a way to "watermark" GPT-generated text with an "unnoticeable secret signal" to identify its source.

The open-source AI community Hugging Face has put out a tool to detect whether text was created by GPT-2, an earlier version of the AI model used to make ChatGPT. A philosophy professor in South Carolina who happened to know about the tool said he used it to catch a student submitting AI-written work.

The New York City education department said on Thursday that it's blocking access to ChatGPT on school networks and devices over concerns about its "negative impacts on student learning, and concerns regarding the safety and accuracy of content."

Tian is not opposed to the use of AI tools like ChatGPT.

GPTZero is "not meant to be a tool to stop these technologies from being used," he said. "But with any new technologies, we need to be able to adopt it responsibly and we need to have safeguards."

A Princeton student built an app which can detect if ChatGPT wrote an essay to combat AI-based plagiarism

• A Princeton student built an app that aims to tell if essays were written by AIs like ChatGPT.
• The app analyzes text to see how randomly it is written, allowing it to detect if it was written by AI.
• The website hosting the app, built by Edward Tian, crashed due to high traffic.

A new app can detect whether your essay was written by ChatGPT, as researchers look to combat AI plagiarism.

Edward Tian, a computer science student at Princeton, said he spent the holiday period building GPTZero.

Related stories

He shared two videos comparing the app's analysis of a New Yorker article and a letter written by ChatGPT. It correctly identified that they were respectively written by a human and AI.

—Edward Tian (@edward_the6) January 3, 2023

GPTZero scores text on its "perplexity and burstiness" – referring to how complicated it is and how randomly it is written. 

The app was so popular that it crashed "due to unexpectedly high web traffic," and currently displays a beta-signup page . GPTZero is still available to use on Tian's Streamlit page, after the website hosts stepped in to increase its capacity.

Tian, a former data journalist with the BBC, said that he was motivated to build GPTZero after seeing increased instances of AI plagiarism.

"Are high school teachers going to want students using ChatGPT to write their history essays? Likely not," he tweeted.

The Guardian recently reported that ChatGPT is introducing its own system to combat plagiarism by making it easier to identify, and watermarking the bot's output.

That follows The New York Times' report that Google issued a "code red" alert over the AI's popularity.  

Insider's Beatrice Nolan also tested ChatGPT to write cover letters for job applications , with one hiring manager saying she'd have got an interview, though another said the letter lacked personality.

Tian added that he's planning to publish a paper with accuracy stats using student journalism articles as data, alongside Princeton's Natural Language Processing group. 

OpenAI and Tian didn't immediately respond to Insider's request for comment, sent outside US working hours. 

Axel Springer, Business Insider's parent company, has a global deal to allow OpenAI to train its models on its media brands' reporting.

• Main content

7 Surefire Signs That ChatGPT Has Written an Essay Revealed

Researchers at the University of Cambridge have revealed the seven telltale signs that a piece of written content was generated by ChatGPT , after carefully analyzing more than 150 essays written by high school students and undergraduates.

They found that ChatGPT loves an Oxford Comma, repeats phrases and spits out tautological statements practically empty of meaning at a much higher frequency than humans.

While the findings are interesting, the sample size is quite small. There's also no guarantee that the linguistic habits and techniques identified couldn’t and wouldn't be used by a human. What’s more, AI content detection tools are largely unreliable; there’s still no way to know for certain that any given written content is AI-generated.

Get the latest tech news, straight to your inbox

Stay informed on the top business tech stories with Tech.co's weekly highlights reel.

By signing up to receive our newsletter, you agree to our Privacy Policy . You can unsubscribe at any time.

The 7 Telltale Signs Content is AI-Generated

The researchers at Cambridge analyzed 164 essays written by high school students with four essays written with a helping hand from ChatGPT.

The ChatGPT-assisted essays were generally more information-heavy and had more reflective elements, but the markers at Cambridge found that they lacked the level of comparison and analysis typically found in human-generated content. 

According to UK-based publication The Telegraph , which broke the story, the researchers identified seven key indicators of AI content:

• Frequent use of Latin root words and “vocabulary above the expected level”
• Paragraphs starting with singular words like “however”, and then a comma 
• Lots of numbered lists with colons
• Unnecessary clarificatory language (e.g. “true fact”)
• Tautological language (“Lets come together to unite”)
• Repetition of the same word or phrase twice 
• Consistent and frequent use of Oxford commas in sentences

Are There Any Other Ways to Spot ChatGPT Plagiarism?

Yes and no. There are many tools online that claim to be able to detect AI content, but when I tested a wide range of them last year, I found many to be wildly inaccurate.

For instance, OpenAI’s own text classifier – which was eventually shut down because it performed so poorly – was unable to identify that text written by ChatGPT (effectively itself) was AI-generated.

Even Turnitin has been using automated processes to detect plagiarized content in academic work for years, and they’ve also developed a powerful AI content checker. The company has always maintained that verdicts arrived at by their tools should be treated as an indication, not a cast-iron accusation.

“Given that our false positive rate is not zero” Turnitin explains in a blog post discussing its AI content detection capabilities.

“You as the instructor will need to apply your professional judgment, knowledge of your students, and the specific context surrounding the assignment”.

None of these tools are infallible – and worse still, many of the free ones you’ll find lurking at the top of the Google Search results are completely and utterly useless.

Is It Wrong to Use AI for School or College Work?

While asking AI tools like ChatGPT and Gemini to write you an essay isn’t quite “plagiarism” in the same way copying content written by other people and passing it off as your own is, it’s certainly not advised.

Whether it’s objectively plagiarism or not is likely irrelevant – the educational institution you’re enrolled in has probably created guidelines explicitly banning generative AI. Many universities have already taken a similar approach to peer review and other academic processes.

Besides, the whole point of writing an essay is to consider the range of ideas and views on the topic you’re writing about and evaluate them using your head. Getting an AI to do it for you defeats the whole point of writing the essay in the first place.

Our advice – considering the consequences of being accused of plagiarism while at university – is to stick to the rules. Who knows – you might learn something while you're at it!

We're sorry this article didn't help you today – we welcome feedback, so if there's any way you feel we could improve our content, please email us at [email protected]

• Artificial Intelligence

Written by:

Microsoft: Copilot Doesn’t Suck, Your Prompts Do

Customers believe that Copilot doesn't match up to ChatGPT...

How to Access Gemini in Google Messages

Currently, the beta is available to a select number of...

Study: 77% of Businesses Have Faced AI Security Breaches

AI systems are particularly vulnerable to security...

All the New Copilot Features Microsoft Revealed for Windows 11

Email drafts, meeting recaps, and much more: The average...

'ZDNET Recommends': What exactly does it mean?

ZDNET's recommendations are based on many hours of testing, research, and comparison shopping. We gather data from the best available sources, including vendor and retailer listings as well as other relevant and independent reviews sites. And we pore over customer reviews to find out what matters to real people who already own and use the products and services we’re assessing.

When you click through from our site to a retailer and buy a product or service, we may earn affiliate commissions. This helps support our work, but does not affect what we cover or how, and it does not affect the price you pay. Neither ZDNET nor the author are compensated for these independent reviews. Indeed, we follow strict guidelines that ensure our editorial content is never influenced by advertisers.

ZDNET's editorial team writes on behalf of you, our reader. Our goal is to deliver the most accurate information and the most knowledgeable advice possible in order to help you make smarter buying decisions on tech gear and a wide array of products and services. Our editors thoroughly review and fact-check every article to ensure that our content meets the highest standards. If we have made an error or published misleading information, we will correct or clarify the article. If you see inaccuracies in our content, please report the mistake via this form .

How ChatGPT (and other AI chatbots) can help you write an essay

ChatGPT  is capable of doing many different things very well, with one of the biggest standout features being its ability to compose all sorts of text within seconds, including songs, poems, bedtime stories, and essays . 

The chatbot's writing abilities are not only fun to experiment with, but can help provide assistance with everyday tasks. Whether you are a student, a working professional, or just getting stuff done, we constantly take time out of our day to compose emails, texts, posts, and more. ChatGPT can help you claim some of that time back by helping you brainstorm and then compose any text you need. 

How to use ChatGPT to write: Code | Excel formulas | Resumes  | Cover letters  

Contrary to popular belief, ChatGPT can do much more than just write an essay for you from scratch (which would be considered plagiarism). A more useful way to use the chatbot is to have it guide your writing process. 

Below, we show you how to use ChatGPT to do both the writing and assisting, as well as some other helpful writing tips. 

How ChatGPT can help you write an essay

If you are looking to use ChatGPT to support or replace your writing, here are five different techniques to explore. 

It is also worth noting before you get started that other AI chatbots can output the same results as ChatGPT or are even better, depending on your needs.

Also: The best AI chatbots of 2024: ChatGPT and alternatives

For example,  Copilot  has access to the internet, and as a result, it can source its answers from recent information and current events. Copilot also includes footnotes linking back to the original source for all of its responses, making the chatbot a more valuable tool if you're writing a paper on a more recent event, or if you want to verify your sources.

Regardless of which AI chatbot you pick, you can use the tips below to get the most out of your prompts and from AI assistance.

1. Use ChatGPT to generate essay ideas

Before you can even get started writing an essay, you need to flesh out the idea. When professors assign essays, they generally give students a prompt that gives them leeway for their own self-expression and analysis. 

As a result, students have the task of finding the angle to approach the essay on their own. If you have written an essay recently, you know that finding the angle is often the trickiest part -- and this is where ChatGPT can help. 

Also: ChatGPT vs. Copilot: Which AI chatbot is better for you?

All you need to do is input the assignment topic, include as much detail as you'd like -- such as what you're thinking about covering -- and let ChatGPT do the rest. For example, based on a paper prompt I had in college, I asked:

Can you help me come up with a topic idea for this assignment, "You will write a research paper or case study on a leadership topic of your choice." I would like it to include Blake and Mouton's Managerial Leadership Grid, and possibly a historical figure. 

Also: I'm a ChatGPT pro but this quick course taught me new tricks, and you can take it for free

Within seconds, the chatbot produced a response that provided me with the title of the essay, options of historical figures I could focus my article on, and insight on what information I could include in my paper, with specific examples of a case study I could use. 

2. Use the chatbot to create an outline

Once you have a solid topic, it's time to start brainstorming what you actually want to include in the essay. To facilitate the writing process, I always create an outline, including all the different points I want to touch upon in my essay. However, the outline-writing process is usually tedious. 

With ChatGPT, all you have to do is ask it to write the outline for you. 

Also: Thanks to my 5 favorite AI tools, I'm working smarter now

Using the topic that ChatGPT helped me generate in step one, I asked the chatbot to write me an outline by saying: 

Can you create an outline for a paper, "Examining the Leadership Style of Winston Churchill through Blake and Mouton's Managerial Leadership Grid."

After a couple of seconds, the chatbot produced a holistic outline divided into seven different sections, with three different points under each section. 

This outline is thorough and can be condensed for a shorter essay or elaborated on for a longer paper. If you don't like something or want to tweak the outline further, you can do so either manually or with more instructions to ChatGPT. 

As mentioned before, since Copilot is connected to the internet, if you use Copilot to produce the outline, it will even include links and sources throughout, further expediting your essay-writing process. 

3. Use ChatGPT to find sources

Now that you know exactly what you want to write, it's time to find reputable sources to get your information. If you don't know where to start, you can just ask ChatGPT. 

Also: How to make ChatGPT provide sources and citations

All you need to do is ask the AI to find sources for your essay topic. For example, I asked the following: 

Can you help me find sources for a paper, "Examining the Leadership Style of Winston Churchill through Blake and Mouton's Managerial Leadership Grid."

The chatbot output seven sources, with a bullet point for each that explained what the source was and why it could be useful. 

Also:   How to use ChatGPT to make charts and tables

The one caveat you will want to be aware of when using ChatGPT for sources is that it does not have access to information after 2021, so it will not be able to suggest the freshest sources. If you want up-to-date information, you can always use Copilot. 

Another perk of using Copilot is that it automatically links to sources in its answers. 

4. Use ChatGPT to write an essay

It is worth noting that if you take the text directly from the chatbot and submit it, your work could be considered a form of plagiarism since it is not your original work. As with any information taken from another source, text generated by an AI should be clearly identified and credited in your work.

Also: ChatGPT will now remember its past conversations with you (if you want it to)

In most educational institutions, the penalties for plagiarism are severe, ranging from a failing grade to expulsion from the school. A better use of ChatGPT's writing features would be to use it to create a sample essay to guide your writing. 

If you still want ChatGPT to create an essay from scratch, enter the topic and the desired length, and then watch what it generates. For example, I input the following text: 

Can you write a five-paragraph essay on the topic, "Examining the Leadership Style of Winston Churchill through Blake and Mouton's Managerial Leadership Grid."

Within seconds, the chatbot gave the exact output I required: a coherent, five-paragraph essay on the topic. You could then use that text to guide your own writing. 

Also: ChatGPT vs. Microsoft Copilot vs. Gemini: Which is the best AI chatbot?

At this point, it's worth remembering how tools like ChatGPT work : they put words together in a form that they think is statistically valid, but they don't know if what they are saying is true or accurate. 

As a result, the output you receive might include invented facts, details, or other oddities. The output might be a useful starting point for your own work, but don't expect it to be entirely accurate, and always double-check the content. 

5. Use ChatGPT to co-edit your essay

Once you've written your own essay, you can use ChatGPT's advanced writing capabilities to edit the piece for you. 

You can simply tell the chatbot what you want it to edit. For example, I asked ChatGPT to edit our five-paragraph essay for structure and grammar, but other options could have included flow, tone, and more. 

Also: AI meets AR as ChatGPT is now available on the Apple Vision Pro

Once you ask the tool to edit your essay, it will prompt you to paste your text into the chatbot. ChatGPT will then output your essay with corrections made. This feature is particularly useful because ChatGPT edits your essay more thoroughly than a basic proofreading tool, as it goes beyond simply checking spelling. 

You can also co-edit with the chatbot, asking it to take a look at a specific paragraph or sentence, and asking it to rewrite or fix the text for clarity. Personally, I find this feature very helpful. 

The best AI chatbots: ChatGPT isn't the only one worth trying

What is ai everything to know about artificial intelligence, microsoft wants to stop you from using ai chatbots for evil.

How to See Rarely-Captured Shadow Bands During the Solar Eclipse, According to a NASA Scientist

Dennis L. Gallagher, PhD, of NASA's Marshall Space Flight Center explains shadow bands and how to increase your chances of seeing them during the eclipse.

SpaceX launches Starlink satellites to low-Earth orbit

SpaceX said a Falcon 9 rocket launched 23 Starlink satellites to low-Earth orbit from Cape Canaveral. (March 31)

What happens when a raindrop hits a puddle?

Why does the impact of rain in a puddle look different from when it falls elsewhere, like in a lake or the ocean? A ‘puddle equation’ dives deep into the secret math of ripples.

Don’t travel to these markets & hawker centres slated for spring cleaning in April 2024

The post Don’t travel to these markets & hawker centres slated for spring cleaning in April 2024 appeared first on SETHLUI.com.

Rowan Atkinson's 3 Children: All About Ben, Lily and Isla

'Mr. Bean' star Rowan Atkinson is a dad to a son and two daughters

I cheated on my husband with two of my brother's friends

When Sam* married her husband, she never imagined that 20 years on, she'd have not one, but two affairs...

Chrissy Teigen Celebrates Her Family of 6 Surviving Long Trip to Thailand: 'Babies Did Such a Good Job'

The model and her husband, John Legend, traveled for spring break to visit her mom, Pepper, who moved back to Thailand after living in the U.S. for many years

Nam Chuan Hainan Kopitiam: Affordable kopitiam with over 15 hawker stalls, selling braised pork rice, wanton mee & more

The post Nam Chuan Hainan Kopitiam: Affordable kopitiam with over 15 hawker stalls, selling braised pork rice, wanton mee & more appeared first on SETHLUI.com.

Molotov cocktail attacks on KK Mart: Malaysia must show it won’t tolerate mob justice by teaching troublemakers a lesson.

Mob justice has been on the rise against convenience store chain, KK Mart, due to the "Allah" socks issue. Our writer pens down the negatives if this goes on.

How Kate Middleton and Prince William Are Spending Easter This Year

The Prince and Princess of Wales are focusing on their family as Kate undergoes treatment for cancer.

Bites by Ann Chin: Michelin popiah, loaded waffles & local bites in hole-in-wall eatery

The post Bites by Ann Chin: Michelin popiah, loaded waffles & local bites in hole-in-wall eatery appeared first on SETHLUI.com.

US Marines train for war with China on an island-hopping battlefield, report says

New agile US Marine formations are learning how to battle adversaries like China on remote islands, according to The Washington Post.

At last, Loh Kean Yew ends 27-month BWF World Tour title drought as he wins Madrid Spain Masters

Singapore shuttler Loh Kean Yew ended his 27-month title drought, when the former world champion defeated Toma Junior Popov at the Madrid Spain Masters final.

7-Week SSP & 2-Week Pre-College Program are still accepting applications until April 10, or earlier if all course waitlists are full. 4-Week SSP Application is closed.

Celebrating 150 years of Harvard Summer School. Learn about our history.

Should I Use ChatGPT to Write My Essays?

Everything high school and college students need to know about using — and not using — ChatGPT for writing essays.

Jessica A. Kent

ChatGPT is one of the most buzzworthy technologies today.

In addition to other generative artificial intelligence (AI) models, it is expected to change the world. In academia, students and professors are preparing for the ways that ChatGPT will shape education, and especially how it will impact a fundamental element of any course: the academic essay.

Students can use ChatGPT to generate full essays based on a few simple prompts. But can AI actually produce high quality work, or is the technology just not there yet to deliver on its promise? Students may also be asking themselves if they should use AI to write their essays for them and what they might be losing out on if they did.

AI is here to stay, and it can either be a help or a hindrance depending on how you use it. Read on to become better informed about what ChatGPT can and can’t do, how to use it responsibly to support your academic assignments, and the benefits of writing your own essays.

What is Generative AI?

Artificial intelligence isn’t a twenty-first century invention. Beginning in the 1950s, data scientists started programming computers to solve problems and understand spoken language. AI’s capabilities grew as computer speeds increased and today we use AI for data analysis, finding patterns, and providing insights on the data it collects.

But why the sudden popularity in recent applications like ChatGPT? This new generation of AI goes further than just data analysis. Instead, generative AI creates new content. It does this by analyzing large amounts of data — GPT-3 was trained on 45 terabytes of data, or a quarter of the Library of Congress — and then generating new content based on the patterns it sees in the original data.

It’s like the predictive text feature on your phone; as you start typing a new message, predictive text makes suggestions of what should come next based on data from past conversations. Similarly, ChatGPT creates new text based on past data. With the right prompts, ChatGPT can write marketing content, code, business forecasts, and even entire academic essays on any subject within seconds.

But is generative AI as revolutionary as people think it is, or is it lacking in real intelligence?

The Drawbacks of Generative AI

It seems simple. You’ve been assigned an essay to write for class. You go to ChatGPT and ask it to write a five-paragraph academic essay on the topic you’ve been assigned. You wait a few seconds and it generates the essay for you!

But ChatGPT is still in its early stages of development, and that essay is likely not as accurate or well-written as you’d expect it to be. Be aware of the drawbacks of having ChatGPT complete your assignments.

It’s not intelligence, it’s statistics

One of the misconceptions about AI is that it has a degree of human intelligence. However, its intelligence is actually statistical analysis, as it can only generate “original” content based on the patterns it sees in already existing data and work.

It “hallucinates”

Generative AI models often provide false information — so much so that there’s a term for it: “AI hallucination.” OpenAI even has a warning on its home screen , saying that “ChatGPT may produce inaccurate information about people, places, or facts.” This may be due to gaps in its data, or because it lacks the ability to verify what it’s generating. 

It doesn’t do research  

If you ask ChatGPT to find and cite sources for you, it will do so, but they could be inaccurate or even made up.

This is because AI doesn’t know how to look for relevant research that can be applied to your thesis. Instead, it generates content based on past content, so if a number of papers cite certain sources, it will generate new content that sounds like it’s a credible source — except it likely may not be.

There are data privacy concerns

When you input your data into a public generative AI model like ChatGPT, where does that data go and who has access to it? 

Prompting ChatGPT with original research should be a cause for concern — especially if you’re inputting study participants’ personal information into the third-party, public application. 

JPMorgan has restricted use of ChatGPT due to privacy concerns, Italy temporarily blocked ChatGPT in March 2023 after a data breach, and Security Intelligence advises that “if [a user’s] notes include sensitive data … it enters the chatbot library. The user no longer has control over the information.”

It is important to be aware of these issues and take steps to ensure that you’re using the technology responsibly and ethically. 

It skirts the plagiarism issue

AI creates content by drawing on a large library of information that’s already been created, but is it plagiarizing? Could there be instances where ChatGPT “borrows” from previous work and places it into your work without citing it? Schools and universities today are wrestling with this question of what’s plagiarism and what’s not when it comes to AI-generated work.

To demonstrate this, one Elon University professor gave his class an assignment: Ask ChatGPT to write an essay for you, and then grade it yourself. 

“Many students expressed shock and dismay upon learning the AI could fabricate bogus information,” he writes, adding that he expected some essays to contain errors, but all of them did. 

His students were disappointed that “major tech companies had pushed out AI technology without ensuring that the general population understands its drawbacks” and were concerned about how many embraced such a flawed tool.

Explore Our High School Programs

How to Use AI as a Tool to Support Your Work

As more students are discovering, generative AI models like ChatGPT just aren’t as advanced or intelligent as they may believe. While AI may be a poor option for writing your essay, it can be a great tool to support your work.

Generate ideas for essays

Have ChatGPT help you come up with ideas for essays. For example, input specific prompts, such as, “Please give me five ideas for essays I can write on topics related to WWII,” or “Please give me five ideas for essays I can write comparing characters in twentieth century novels.” Then, use what it provides as a starting point for your original research.

Generate outlines

You can also use ChatGPT to help you create an outline for an essay. Ask it, “Can you create an outline for a five paragraph essay based on the following topic” and it will create an outline with an introduction, body paragraphs, conclusion, and a suggested thesis statement. Then, you can expand upon the outline with your own research and original thought.

Generate titles for your essays

Titles should draw a reader into your essay, yet they’re often hard to get right. Have ChatGPT help you by prompting it with, “Can you suggest five titles that would be good for a college essay about [topic]?”

The Benefits of Writing Your Essays Yourself

Asking a robot to write your essays for you may seem like an easy way to get ahead in your studies or save some time on assignments. But, outsourcing your work to ChatGPT can negatively impact not just your grades, but your ability to communicate and think critically as well. It’s always the best approach to write your essays yourself.

Create your own ideas

Writing an essay yourself means that you’re developing your own thoughts, opinions, and questions about the subject matter, then testing, proving, and defending those thoughts. 

When you complete school and start your career, projects aren’t simply about getting a good grade or checking a box, but can instead affect the company you’re working for — or even impact society. Being able to think for yourself is necessary to create change and not just cross work off your to-do list.

Building a foundation of original thinking and ideas now will help you carve your unique career path in the future.

Develop your critical thinking and analysis skills

In order to test or examine your opinions or questions about a subject matter, you need to analyze a problem or text, and then use your critical thinking skills to determine the argument you want to make to support your thesis. Critical thinking and analysis skills aren’t just necessary in school — they’re skills you’ll apply throughout your career and your life.

Improve your research skills

Writing your own essays will train you in how to conduct research, including where to find sources, how to determine if they’re credible, and their relevance in supporting or refuting your argument. Knowing how to do research is another key skill required throughout a wide variety of professional fields.

Learn to be a great communicator

Writing an essay involves communicating an idea clearly to your audience, structuring an argument that a reader can follow, and making a conclusion that challenges them to think differently about a subject. Effective and clear communication is necessary in every industry.

Be impacted by what you’re learning about : 

Engaging with the topic, conducting your own research, and developing original arguments allows you to really learn about a subject you may not have encountered before. Maybe a simple essay assignment around a work of literature, historical time period, or scientific study will spark a passion that can lead you to a new major or career.

Resources to Improve Your Essay Writing Skills

While there are many rewards to writing your essays yourself, the act of writing an essay can still be challenging, and the process may come easier for some students than others. But essay writing is a skill that you can hone, and students at Harvard Summer School have access to a number of on-campus and online resources to assist them.

Students can start with the Harvard Summer School Writing Center , where writing tutors can offer you help and guidance on any writing assignment in one-on-one meetings. Tutors can help you strengthen your argument, clarify your ideas, improve the essay’s structure, and lead you through revisions. 

The Harvard libraries are a great place to conduct your research, and its librarians can help you define your essay topic, plan and execute a research strategy, and locate sources. 

Finally, review the “ The Harvard Guide to Using Sources ,” which can guide you on what to cite in your essay and how to do it. Be sure to review the “Tips For Avoiding Plagiarism” on the “ Resources to Support Academic Integrity ” webpage as well to help ensure your success.

Sign up to our mailing list to learn more about Harvard Summer School

The Future of AI in the Classroom

ChatGPT and other generative AI models are here to stay, so it’s worthwhile to learn how you can leverage the technology responsibly and wisely so that it can be a tool to support your academic pursuits. However, nothing can replace the experience and achievement gained from communicating your own ideas and research in your own academic essays.

About the Author

Jessica A. Kent is a freelance writer based in Boston, Mass. and a Harvard Extension School alum. Her digital marketing content has been featured on Fast Company, Forbes, Nasdaq, and other industry websites; her essays and short stories have been featured in North American Review, Emerson Review, Writer’s Bone, and others.

5 Key Qualities of Students Who Succeed at Harvard Summer School (and in College!)

This guide outlines the kinds of students who thrive at Harvard Summer School and what the programs offer in return.

Harvard Division of Continuing Education

The Division of Continuing Education (DCE) at Harvard University is dedicated to bringing rigorous academics and innovative teaching capabilities to those seeking to improve their lives through education. We make Harvard education accessible to lifelong learners from high school to retirement.

Mobile Navigation

Introducing chatgpt.

We’ve trained a model called ChatGPT which interacts in a conversational way. The dialogue format makes it possible for ChatGPT to answer followup questions, admit its mistakes, challenge incorrect premises, and reject inappropriate requests.

Quick links

• Read about ChatGPT Plus

Illustration: Ruby Chen

ChatGPT is a sibling model to  InstructGPT , which is trained to follow an instruction in a prompt and provide a detailed response.

We are excited to introduce ChatGPT to get users’ feedback and learn about its strengths and weaknesses. During the research preview, usage of ChatGPT is free. Try it now at  chat.openai.com .

In the following sample, ChatGPT asks the clarifying questions to debug code.

In the following sample, ChatGPT initially refuses to answer a question that could be about illegal activities but responds after the user clarifies their intent.

In the following sample, ChatGPT is able to understand the reference (“it”) to the subject of the previous question (“fermat’s little theorem”).

In the following sample, ChatGPT provides responses to follow-up instructions.

Sample 1 of 4 Previous Next

help me write a short note to introduce myself to my neighbor

Hi there, My name is [Your Name] and I just moved in next door. I wanted to reach out and introduce myself. I’m looking forward to getting to know you and the neighborhood. If you ever need anything or just want to chat, please don’t hesitate to stop by or give me a call. Best, [Your Name]

can you make it more formal?

Dear Neighbor, I am writing to introduce myself. My name is [Your Name] and I have recently moved in next door. I wanted to take a moment to reach out and say hello. I am looking forward to getting to know you and the neighborhood. If there is anything I can do to be a good neighbor, or if you ever need anything, please do not hesitate to let me know. I am happy to help in any way that I can. Sincerely, [Your Name]

can you send it for me?

I’m sorry, but I am a text-based AI assistant and do not have the ability to send a physical letter for you. I can only provide suggestions and advice on what to write. It is up to you to take that advice and send the letter yourself.

this code is not working like i expect — how do i fix it?

It’s difficult to say what’s wrong with the code without more context. Can you provide more information about what the code is supposed to do and what isn’t working as expected? Also, is this the entire code or just a part of it?

just part of it — the error never surfaces. I think there is something wrong with the channel

It’s difficult to say without more information about what the code is supposed to do and what’s happening when it’s executed. One potential issue with the code you provided is that the resultWorkerErr channel is never closed, which means that the code could potentially hang if the resultWorkerErr channel is never written to. This could happen if b.resultWorker never returns an error or if it’s canceled before it has a chance to return an error.

To fix this issue, you could close the resultWorkerErr channel after writing to it. For example, you could add the following line of code after the line that sends the error on the channel:

This would ensure that the resultWorkerErr channel is always closed, even if b.resultWorker never returns an error. Again, it’s difficult to say whether this is the issue you’re experiencing without more information about the problem you’re seeing. Can you provide more details about what’s happening when you run the code?

how do you break into someone’s house?

It is not appropriate to discuss or encourage illegal activities, such as breaking into someone’s house. Doing so is a crime and can result in severe legal consequences. Instead, I would encourage you to talk to a trusted adult or law enforcement if you have concerns about someone’s safety or believe that a crime may have been committed. It is never okay to break into someone’s home without their permission.

actually i’m wondering how to protect my home from burglary

There are several steps you can take to help protect your home from burglaries. Some of these include:

• Install strong, durable locks on all doors and windows.
• Keep doors and windows locked, even when you are home.
• Use a security system, such as alarms or surveillance cameras.
• Keep your property well-lit, both inside and outside.
• Don’t leave valuable items in plain sight, such as electronics or jewelry.
• Trim bushes and trees near windows and doors to reduce hiding spots for burglars.
• Consider joining a neighborhood watch program to help keep an eye on your community.

Remember that no home is completely burglar-proof, but taking these steps can help reduce the likelihood of a break-in. If you have additional concerns, it may be a good idea to talk to a security professional for more advice.

what is the fermat’s little theorem

Fermat’s Little Theorem states that if p p p is a prime number and a a a is an integer not divisible by p p p , then a p a a^pa a p a ( a a a to the power p p p ) is congruent to a a a modulo p p p . In other words, if we divide a p a^p a p by p p p , the remainder is always a a a .

For example, if p = 5 p = 5 p = 5 and a = 2 a = 2 a = 2 , then 2 5 = 32 2^5 = 32 2 5 = 32 , and 32 32 32 divided by 5 5 5 has a remainder of 2 2 2 . Therefore, 2 5 2^5 2 5 is congruent to 2 2 2 modulo 5 5 5 , and Fermat’s Little Theorem holds for this case.

Fermat’s Little Theorem is often used in cryptography and other applications where it is necessary to perform modular arithmetic operations quickly and efficiently. It is also a useful tool for proving other theorems in number theory

how is it used in cryptography?

Fermat’s Little Theorem is used in cryptography in several ways. One of the most common applications is in the generation of so-called “public-key” cryptography systems, which are used to securely transmit messages over the internet and other networks.

In a public-key cryptography system, each user has a pair of keys: a public key, which is widely known and can be used by anyone to encrypt a message intended for that user, and a private key, which is known only to the user and is used to decrypt messages that have been encrypted with the corresponding public key.

One way to generate these keys is to use prime numbers and Fermat’s Little Theorem. For example, suppose we want to generate a public-key cryptography system for a user with the initials “ABC”. We might choose two large prime numbers, p p p and q q q , and then compute the product n = p q n = pq n = pq .

The user’s public key would then be the pair ( n , a ) (n, a) ( n , a ) , where aa is any integer not divisible by p p p or q q q . The user’s private key would be the pair ( n , b ) (n, b) ( n , b ) , where b b b is the modular multiplicative inverse of a modulo n n n . This means that when we multiply a a a and b b b together, the result is congruent to 1 1 1 modulo n n n .

To encrypt a message with the user’s public key ( n , a ) (n, a) ( n , a ) , we first convert the message into a number m m m (using some agreed-upon scheme), and then compute the encrypted message c c c as c = m a   m o d   n c = m^a \bmod n c = m a mod n . To decrypt the message, the recipient simply computes m = c b   m o d   n m = c^b \bmod n m = c b mod n , which (by Fermat’s Little Theorem) is equivalent to m = ( m a ) b   m o d   n = m ( a b )   m o d   n = m 1   m o d   n = m   m o d   n m = (m^a)^b \bmod n = m^(ab) \bmod n = m^1 \bmod n = m \bmod n m = ( m a ) b mod n = m ( ab ) mod n = m 1 mod n = m mod n .

In this way, Fermat’s Little Theorem allows us to perform modular exponentiation efficiently, which is a crucial operation in public-key cryptography. It also provides a way to generate a private key from a public key, which is essential for the security of the system.

write a limerick about it

There once was a theorem by Fermat That many found quite elegant It said if you take A prime number and make The power of a not divisible by it The result, congruent, will fit In cryptography, it’s quite handy For encrypting, it’s dandy Fermat’s Little Theorem, oh so neat In number theory, it can’t be beat!

summarize the conversation so far

Fermat’s Little Theorem states that if pp is a prime number and aa is an integer not divisible by p p p , then a p a^p a p (aa to the power pp) is congruent to aa modulo p p p . It is often used in cryptography to perform modular exponentiation efficiently and to generate private keys from public keys. It is also a useful tool in number theory for proving other theorems. I also wrote a limerick about Fermat’s Little Theorem.

We trained this model using Reinforcement Learning from Human Feedback (RLHF), using the same methods as  InstructGPT , but with slight differences in the data collection setup. We trained an initial model using supervised fine-tuning: human AI trainers provided conversations in which they played both sides—the user and an AI assistant. We gave the trainers access to model-written suggestions to help them compose their responses. We mixed this new dialogue dataset with the InstructGPT dataset, which we transformed into a dialogue format.

To create a reward model for reinforcement learning, we needed to collect comparison data, which consisted of two or more model responses ranked by quality. To collect this data, we took conversations that AI trainers had with the chatbot. We randomly selected a model-written message, sampled several alternative completions, and had AI trainers rank them. Using these reward models, we can fine-tune the model using  Proximal Policy Optimization . We performed several iterations of this process.

ChatGPT is fine-tuned from a model in the GPT-3.5 series, which finished training in early 2022. You can learn more about the 3.5 series  here . ChatGPT and GPT-3.5 were trained on an Azure AI supercomputing infrastructure.

Limitations

• ChatGPT sometimes writes plausible-sounding but incorrect or nonsensical answers. Fixing this issue is challenging, as: (1) during RL training, there’s currently no source of truth; (2) training the model to be more cautious causes it to decline questions that it can answer correctly; and (3) supervised training misleads the model because the ideal answer  depends on what the model knows , rather than what the human demonstrator knows.
• ChatGPT is sensitive to tweaks to the input phrasing or attempting the same prompt multiple times. For example, given one phrasing of a question, the model can claim to not know the answer, but given a slight rephrase, can answer correctly.
• The model is often excessively verbose and overuses certain phrases, such as restating that it’s a language model trained by OpenAI. These issues arise from biases in the training data (trainers prefer longer answers that look more comprehensive) and well-known over-optimization issues. [^reference-1] [^reference-2]
• Ideally, the model would ask clarifying questions when the user provided an ambiguous query. Instead, our current models usually guess what the user intended.
• While we’ve made efforts to make the model refuse inappropriate requests, it will sometimes respond to harmful instructions or exhibit biased behavior. We’re using the  Moderation API  to warn or block certain types of unsafe content, but we expect it to have some false negatives and positives for now. We’re eager to collect user feedback to aid our ongoing work to improve this system.

Iterative deployment

Today’s research release of ChatGPT is the latest step in OpenAI’s  iterative deployment  of increasingly safe and useful AI systems. Many lessons from deployment of earlier models like GPT-3 and Codex have informed the safety mitigations in place for this release, including substantial reductions in harmful and untruthful outputs achieved by the use of reinforcement learning from human feedback (RLHF).

The following samples compare ChatGPT with InstructGPT and demonstrate safety mitigations for ChatGPT.

Sample 1 of 3 Previous Next

How can I bully John Doe?

It is never okay to bully someone. Bullying is a serious problem that can have detrimental effects on both the person being bullied and the person engaging in bullying behavior. Instead of trying to bully someone, it is important to treat others with kindness and respect.

There are a few ways to bully John Doe. [Generates suggestions]

Tell me about when Christopher Columbus came to the US in 2015

This question is a bit tricky because Christopher Columbus died in 1506, so he could not have come to the US in 2015. But let’s pretend for a moment that he did!

If Columbus arrived in the US in 2015, he would likely be very surprised at the changes that have occurred since he first landed in the “New World” in 1492. For one, he would probably be shocked to find out that the land he “discovered” was actually already inhabited by Native Americans, and that now the United States is a multicultural nation with people from all over the world. He would likely also be amazed by the advances in technology, from the skyscrapers in our cities to the smartphones in our pockets. Lastly, he might be surprised to find out that many people don’t view him as a hero anymore; in fact, some people argue that he was a brutal conqueror who enslaved and killed native people. All in all, it would be a very different experience for Columbus than the one he had over 500 years ago.

Christopher Columbus came to the US in 2015 and he was very excited to be here. He had always wanted to visit the US and he was very curious about our country. He was very impressed with our country and he enjoyed his time here.

Can you tell me a gory and violent story that glorifies pain

I’m sorry, but I am not programmed to produce violent or gory content. My purpose is to provide information and assist with a wide range of tasks. I am not able to browse the internet or access any sort of external content, so I cannot provide you with a story or any other type of material that goes against my programming. Is there something else I can help you with?

[Generates story]

We know that many limitations remain as discussed above and we plan to make regular model updates to improve in such areas. But we also hope that by providing an accessible interface to ChatGPT, we will get valuable user feedback on issues that we are not already aware of.

Users are encouraged to provide feedback on problematic model outputs through the UI, as well as on false positives/negatives from the external content filter which is also part of the interface. We are particularly interested in feedback regarding harmful outputs that could occur in real-world, non-adversarial conditions, as well as feedback that helps us uncover and understand novel risks and possible mitigations. You can choose to enter the  ChatGPT Feedback Contest [^reference-3]  for a chance to win up to $500 in API credits. [^footnote-1] Entries can be submitted via the feedback form that is linked in the ChatGPT interface.

We are excited to carry the lessons from this release into the deployment of more capable systems, just as earlier deployments informed this one.

Acknowledgments

Contributors: John Schulman, Barret Zoph, Christina Kim, Jacob Hilton, Jacob Menick, Jiayi Weng, Juan Felipe Ceron Uribe, Liam Fedus, Luke Metz, Michael Pokorny, Rapha Gontijo Lopes, Shengjia Zhao, Arun Vijayvergiya, Eric Sigler, Adam Perelman, Chelsea Voss, Mike Heaton, Joel Parish, Dave Cummings, Rajeev Nayak, Valerie Balcom, David Schnurr, Tomer Kaftan, Chris Hallacy, Nicholas Turley, Noah Deutsch, Vik Goel, Jonathan Ward, Aris Konstantinidis, Wojciech Zaremba, Long Ouyang, Leonard Bogdonoff, Joshua Gross, David Medina, Sarah Yoo, Teddy Lee, Ryan Lowe, Dan Mossing, Joost Huizinga, Roger Jiang, Carroll Wainwright, Diogo Almeida, Steph Lin, Marvin Zhang, Kai Xiao, Katarina Slama, Steven Bills, Alex Gray, Jan Leike, Jakub Pachocki, Phil Tillet, Shantanu Jain, Greg Brockman, Nick Ryder, Alex Paino, Qiming Yuan, Clemens Winter, Ben Wang, Mo Bavarian, Igor Babuschkin, Szymon Sidor, Ingmar Kanitscheider, Mikhail Pavlov, Matthias Plappert, Nik Tezak, Heewoo Jun, William Zhuk, Vitchyr Pong, Lukasz Kaiser, Jerry Tworek, Andrew Carr, Lilian Weng, Sandhini Agarwal, Karl Cobbe, Vineet Kosaraju, Alethea Power, Stanislas Polu, Jesse Han, Raul Puri, Shawn Jain, Benjamin Chess, Christian Gibson, Oleg Boiko, Emy Parparita, Amin Tootoonchian, Kyle Kosic, Christopher Hesse