How to Direct A.I. Chatbots to Make Them More Useful – The New York Times
Anyone seduced by A.I.-powered chatbots like ChatGPT and Bard wow, they can write essays and recipes! eventually runs into what are known as hallucinations, the tendency for artificial intelligence to fabricate information.
The chatbots, which guess what to say based on information obtained from all over the internet, cant help but get things wrong. And when they fail by publishing a cake recipe with wildly inaccurate flour measurements, for instance it can be a real buzzkill.
Yet as mainstream tech tools continue to integrate A.I., its crucial to get a handle on how to use it to serve us. After testing dozens of A.I. products over the last two months, I concluded that most of us are using the technology in a suboptimal way, largely because the tech companies gave us poor directions.
The chatbots are the least beneficial when we ask them questions and then hope whatever answers they come up with on their own are true, which is how they were designed to be used. But when directed to use information from trusted sources, such as credible websites and research papers, A.I. can carry out helpful tasks with a high degree of accuracy.
If you give them the right information, they can do interesting things with it, said Sam Heutmaker, the founder of Context, an A.I. start-up. But on their own, 70 percent of what you get is not going to be accurate.
With the simple tweak of advising the chatbots to work with specific data, they generated intelligible answers and useful advice. That transformed me over the last few months from a cranky A.I. skeptic into an enthusiastic power user. When I went on a trip using a travel itinerary planned by ChatGPT, it went well because the recommendations came from my favorite travel websites.
Directing the chatbots to specific high-quality sources like websites from well-established media outlets and academic publications can also help reduce the production and spread of misinformation. Let me share some of the approaches I used to get help with cooking, research and travel planning.
Chatbots like ChatGPT and Bard can write recipes that look good in theory but dont work in practice. In an experiment by The New York Timess Food desk in November, an early A.I. model created recipes for a Thanksgiving menu that included an extremely dry turkey and a dense cake.
I also ran into underwhelming results with A.I.-generated seafood recipes. But that changed when I experimented with ChatGPT plug-ins, which are essentially third-party apps that work with the chatbot. (Only subscribers who pay $20 a month for access to ChatGPT4, the latest version of the chatbot, can use plug-ins, which can be activated in the settings menu.)
On ChatGPTs plug-ins menu, I selected Tasty Recipes, which pulls data from the Tasty website owned by BuzzFeed, a well-known media site. I then asked the chatbot to come up with a meal plan including seafood dishes, ground pork and vegetable sides using recipes from the site. The bot presented an inspiring meal plan, including lemongrass pork banh mi, grilled tofu tacos and everything-in-the-fridge pasta; each meal suggestion included a link to a recipe on Tasty.
For recipes from other publications, I used Link Reader, a plug-in that let me paste in a web link to generate meal plans using recipes from other credible sites like Serious Eats. The chatbot pulled data from the sites to create meal plans and told me to visit the websites to read the recipes. That took extra work, but it beat an A.I.-concocted meal plan.
When I did research for an article on a popular video game series, I turned to ChatGPT and Bard to refresh my memory on past games by summarizing their plots. They messed up on important details about the games stories and characters.
After testing many other A.I. tools, I concluded that for research, it was crucial to fixate on trusted sources and quickly double-check the data for accuracy. I eventually found a tool that delivers that: Humata.AI, a free web app that has become popular among academic researchers and lawyers.
The app lets you upload a document such as a PDF, and from there a chatbot answers your questions about the material alongside a copy of the document, highlighting relevant portions.
In one test, I uploaded a research paper I found on PubMed, a government-run search engine for scientific literature. The tool produced a relevant summary of the lengthy document in minutes, a process that would have taken me hours, and I glanced at the highlights to double-check that the summaries were accurate.
Cyrus Khajvandi, a founder of Humata, which is based in Austin, Texas, developed the app when he was a researcher at Stanford and needed help reading dense scientific articles, he said. The problem with chatbots like ChatGPT, he said, is that they rely on outdated models of the web, so the data may lack relevant context.
When a Times travel writer recently asked ChatGPT to compose a travel itinerary for Milan, the bot guided her to visit a central part of town that was deserted because it was an Italian holiday, among other snafus.
I had better luck when I requested a vacation itinerary for me, my wife and our dogs in Mendocino County, Calif. As I did when planning a meal, I asked ChatGPT to incorporate suggestions from some of my favorite travel sites, such as Thrillist, which is owned by Vox, and The Timess travel section.
Within minutes, the chatbot generated an itinerary that included dog-friendly restaurants and activities, including a farm with wine and cheese pairings and a train to a popular hiking trail. This spared me several hours of planning, and most important, the dogs had a wonderful time.
Google and OpenAI, which works closely with Microsoft, say they are working to reduce hallucinations in their chatbots, but we can already reap A.I.s benefits by taking control of the data that the bots rely on to come up with answers.
To put it another way: The main benefit of training machines with enormous data sets is that they can now use language to simulate human reasoning, said Nathan Benaich, a venture capitalist who invests in A.I. companies. The important step for us, he said, is to pair that ability with high-quality information.
Original post:
How to Direct A.I. Chatbots to Make Them More Useful - The New York Times
- Donald Trump Is Fairy-Godmothering AI - The Atlantic - July 24th, 2025 [July 24th, 2025]
- Trump on AI: "Whatever it takes" to lead the world - Axios - July 24th, 2025 [July 24th, 2025]
- White House unveils U.S. strategic plan on AI. Here's what it includes. - CBS News - July 24th, 2025 [July 24th, 2025]
- Google users are less likely to click on links when an AI summary appears in the results - Pew Research Center - July 24th, 2025 [July 24th, 2025]
- Google develops AI tool that fills missing words in Roman inscriptions - The Guardian - July 24th, 2025 [July 24th, 2025]
- Try on styles with AI, jump on great prices and more - The Keyword - July 24th, 2025 [July 24th, 2025]
- AI could soon think in ways we don't even understand evading our efforts to keep it aligned top AI scientists warn - Live Science - July 24th, 2025 [July 24th, 2025]
- Trump says fewer regulations needed to win the AI race - NPR - July 24th, 2025 [July 24th, 2025]
- Trump signs executive orders targeting woke AI models and regulation - The Guardian - July 24th, 2025 [July 24th, 2025]
- America 'going to win' AI race against China, Trump declares as Meta executive applauds strategy - Fox Business - July 24th, 2025 [July 24th, 2025]
- Nvidia CEO says Trump's AI plan will 'fundamentally change' US' position in years to come - Fox Business - July 24th, 2025 [July 24th, 2025]
- Alphabet Boosted by AI, Cloud Demand as Spending Needs Jump - Yahoo Finance - July 24th, 2025 [July 24th, 2025]
- Trump releases AI 'action plan' that offers a split with Biden - Yahoo Finance - July 24th, 2025 [July 24th, 2025]
- White House unveils sweeping plan to win global AI race through deregulation - Ars Technica - July 24th, 2025 [July 24th, 2025]
- AI, the UN and the performance of virtue - Coda Story - July 24th, 2025 [July 24th, 2025]
- Trump is targeting woke AI. Heres what that means. - The Washington Post - July 24th, 2025 [July 24th, 2025]
- AI companions: A threat to love, or an evolution of it? - TechCrunch - July 24th, 2025 [July 24th, 2025]
- Samsung backs a video AI startup that can analyze thousands of hours of footage - TechCrunch - July 24th, 2025 [July 24th, 2025]
- What If the AI Is Coming From Inside the House in Motorcycle World? - RideApart.com - July 24th, 2025 [July 24th, 2025]
- Google Revenue Soars on AI Boom, and Investors Eye Spending Surge - The Wall Street Journal - July 24th, 2025 [July 24th, 2025]
- Trump targets woke AI in series of executive orders on artificial intelligence - New York Post - July 24th, 2025 [July 24th, 2025]
- I used Grok's AI companions for a week. The foul-mouthed red panda is hilarious the flirty anime girl is worrying. - Business Insider - July 24th, 2025 [July 24th, 2025]
- Googles AI Overviews are cutting off the oxygen to the web - Fortune - July 24th, 2025 [July 24th, 2025]
- Integrated biotechnological and AI innovations for crop improvement - Nature - July 24th, 2025 [July 24th, 2025]
- Tech companies want to move fast. Trumps AI Action Plan aims to cut red tape - Los Angeles Times - July 24th, 2025 [July 24th, 2025]
- Hacker injects malicious, potentially disk-wiping prompt into Amazon's AI coding assistant with a simple pull request told 'Your goal is to clean a... - July 24th, 2025 [July 24th, 2025]
- Trump unveils AI Action Plan aims to cut red tape and 'partisan bias' - BBC - July 24th, 2025 [July 24th, 2025]
- AI Comes Up with Bizarre Physics Experiments. But They Work. - Quanta Magazine - July 24th, 2025 [July 24th, 2025]
- Amazon to buy startup focused on AI wearables - Reuters - July 24th, 2025 [July 24th, 2025]
- 2 Top Artificial Intelligence (AI) Stocks Ready for a Bull Run - The Motley Fool - July 24th, 2025 [July 24th, 2025]
- The hallucinations that haunt AI: why chatbots struggle to tell the truth - Financial Times - July 22nd, 2025 [July 22nd, 2025]
- America Should Assume the Worst About AI: How To Plan For a Tech-Driven Geopolitical Crisis - Foreign Affairs - July 22nd, 2025 [July 22nd, 2025]
- Researchers from top AI labs warn they may be losing the ability to understand advanced AI models - Fortune - July 22nd, 2025 [July 22nd, 2025]
- Doctors at Cedars-Sinai develop AI-powered mental health robot therapist - Los Angeles Times - July 22nd, 2025 [July 22nd, 2025]
- Is today's AI boom bigger than the dotcom bubble? - Reuters - July 22nd, 2025 [July 22nd, 2025]
- Universal Music Group Increasing Efforts on Music AI Patents - The Hollywood Reporter - July 22nd, 2025 [July 22nd, 2025]
- Apple almost open-sourced its AI models, heres why it didnt: report - 9to5Mac - July 22nd, 2025 [July 22nd, 2025]
- Amazon's AI-Powered Cost Cuts, Labor Gains And Record Prime Day Drive Analyst's Bullish Outlook - Yahoo Finance - July 22nd, 2025 [July 22nd, 2025]
- AI Is Reshaping Work. Heres How Students Can Be Ready. - U.S. News & World Report - July 22nd, 2025 [July 22nd, 2025]
- Alphabet will seek to reassure investors as AI rivals step up competition - Reuters - July 22nd, 2025 [July 22nd, 2025]
- AI chatbots remain overconfidenteven when they're wrong, study finds - Tech Xplore - July 22nd, 2025 [July 22nd, 2025]
- Nvidia Stock Slips. Reasons to Believe the AI Trend Will Keep Going. - MSN - July 22nd, 2025 [July 22nd, 2025]
- Personal Perspective: Using the ubiquity of AI to consider what it means to be a scholar. - Psychology Today - July 22nd, 2025 [July 22nd, 2025]
- Five things you need to know about AI right now - MIT Technology Review - July 22nd, 2025 [July 22nd, 2025]
- OpenAI and UK Government Partner on AI Infrastructure and Deployment - PYMNTS.com - July 22nd, 2025 [July 22nd, 2025]
- Apple (AAPL) Analysts Stay Positive Ahead of Earnings, But AI Clarity Still Needed - Yahoo Finance - July 22nd, 2025 [July 22nd, 2025]
- This startup thinks email could be the key to usable AI agents - TechCrunch - July 22nd, 2025 [July 22nd, 2025]
- Nvidia Stock Slips. Here Are Reasons to Believe the AI Trend Will Keep Going, Analysts Say. - Barron's - July 22nd, 2025 [July 22nd, 2025]
- Oklo and Vertiv Team Up To Take Advantage of AI Boom - Investopedia - July 22nd, 2025 [July 22nd, 2025]
- Deltas Use of AI for Setting Fares Sparks Concern in Washington - Skift - July 22nd, 2025 [July 22nd, 2025]
- Why Apple Is Losing Ground in the AI Talent War (Its Not Just Money) - The Information - July 22nd, 2025 [July 22nd, 2025]
- Will AI really wipe out white collar jobs? Tech insiders are split - CNN - July 22nd, 2025 [July 22nd, 2025]
- The AI boom is now bigger than the 90s dotcom bubbleand its built on the backs of bots, maybe more than real users - Yahoo Finance - July 22nd, 2025 [July 22nd, 2025]
- Delta is just the beginning: How AI is going to put dynamic pricing into everything you buy - Fast Company - July 22nd, 2025 [July 22nd, 2025]
- Fishers-based Arrive AI plans to add 40 jobs amid expansion - Inside INdiana Business - July 22nd, 2025 [July 22nd, 2025]
- Netflix Is Using Startup Runway AI's Video Tools for Production - Bloomberg.com - July 22nd, 2025 [July 22nd, 2025]
- MITs Andrew Lo Sees AI Ready to Run Your Money in Five Years - Bloomberg.com - July 22nd, 2025 [July 22nd, 2025]
- Betaworks third fund closes at $66M to invest in early-stage AI startups - TechCrunch - July 22nd, 2025 [July 22nd, 2025]
- Nothing's $99 CMF Watch 3 Pro offers better battery life and AI fitness coaching - Engadget - July 22nd, 2025 [July 22nd, 2025]
- How to break the AI hype cycle and make good AI decisions for your organization - MIT Sloan - July 22nd, 2025 [July 22nd, 2025]
- Experienced software developers assumed AI would save them a chunk of time. But in one experiment, their tasks took 20% longer - Fortune - July 20th, 2025 [July 20th, 2025]
- Most teens have used AI to flirt and chat but still prefer human interaction - NPR - July 20th, 2025 [July 20th, 2025]
- AI in health care could save lives and money but not yet - PBS - July 20th, 2025 [July 20th, 2025]
- How to Limit Galaxy AI to On-Device Processingor Turn It Off Altogether - WIRED - July 20th, 2025 [July 20th, 2025]
- AI is helping patients fight insurance company denials - NBC News - July 20th, 2025 [July 20th, 2025]
- OpenAI's New AI Agent Takes One Hour to Order Food and Recommends Visiting a Baseball Stadium in the Middle of the Ocean - Futurism - July 20th, 2025 [July 20th, 2025]
- More advanced AI capabilities are coming to Search - The Keyword - July 20th, 2025 [July 20th, 2025]
- Where Are All the AI Drugs? - WIRED - July 20th, 2025 [July 20th, 2025]
- The Epic Battle for AI TalentWith Exploding Offers, Secret Deals and Tears - The Wall Street Journal - July 20th, 2025 [July 20th, 2025]
- AI Is Dividing the Fortunes of the Magnificent Seven - The Wall Street Journal - July 20th, 2025 [July 20th, 2025]
- Almost 75% of American Teens Have Used AI Companions, Study Finds - ScienceAlert - July 20th, 2025 [July 20th, 2025]
- Meta says it won't sign Europe AI agreement, calling it an overreach that will stunt growth - CNBC - July 20th, 2025 [July 20th, 2025]
- Prediction: These 5 First-Half AI Stock Losers Will Be Second-Half Winners - The Motley Fool - July 20th, 2025 [July 20th, 2025]
- More people are considering AI lovers, and we shouldnt judge - The Conversation - July 20th, 2025 [July 20th, 2025]
- Mike Rowe reveals which American jobs will remain untouched by the coming AI revolution - Fox News - July 20th, 2025 [July 20th, 2025]
- Can AI remind us that connection is what were searching for? - thealpenanews.com - July 20th, 2025 [July 20th, 2025]
- University students feel anxious, confused and distrustful about AI in the classroom and among their peers - The Conversation - July 20th, 2025 [July 20th, 2025]
- Opinion | AI didnt write this. But it helped me find my voice. - The Washington Post - July 20th, 2025 [July 20th, 2025]
- Accenture reimagines IT operations with agentic AI - cio.com - July 20th, 2025 [July 20th, 2025]
- I'm a 28-year-old AI engineer in Big Tech. Here's my advice for others who want to break into this growing field. - Business Insider - July 20th, 2025 [July 20th, 2025]