Technologies

AI Is Bad at Sudoku. It’s Even Worse at Showing Its Work

Researchers did more than ask chatbots to play games. They tested whether AI models could describe their thinking. The results were troubling.

Chatbots are genuinely impressive when you watch them do things they’re good at, like writing a basic email or creating weird, futuristic-looking images. But ask generative AI to solve one of those puzzles in the back of a newspaper, and things can quickly go off the rails.

That’s what researchers at the University of Colorado at Boulder found when they challenged large language models to solve sudoku. And not even the standard 9×9 puzzles. An easier 6×6 puzzle was often beyond the capabilities of an LLM without outside help (in this case, specific puzzle-solving tools).

A more important finding came when the models were asked to show their work. For the most part, they couldn’t. Sometimes they lied. Sometimes they explained things in ways that made no sense. Sometimes they hallucinated and started talking about the weather.

If gen AI tools can’t explain their decisions accurately or transparently, that should cause us to be cautious as we give these things more control over our lives and decisions, said Ashutosh Trivedi, a computer science professor at the University of Colorado at Boulder and one of the authors of the paper published in July in the Findings of the Association for Computational Linguistics.

«We would really like those explanations to be transparent and be reflective of why AI made that decision, and not AI trying to manipulate the human by providing an explanation that a human might like,» Trivedi said.

Don’t miss any of our unbiased tech content and lab-based reviews. Add CNET as a preferred Google source.

The paper is part of a growing body of research into the behavior of large language models. Other recent studies have found, for example, that models hallucinate in part because their training procedures incentivize them to produce results a user will like, rather than what is accurate, or that people who use LLMs to help them write essays are less likely to remember what they wrote. As gen AI becomes more and more a part of our daily lives, the implications of how this technology works and how we behave when using it become hugely important.

When you make a decision, you can try to justify it, or at least explain how you arrived at it. An AI model may not be able to accurately or transparently do the same. Would you trust it?

Why LLMs struggle with sudoku

We’ve seen AI models fail at basic games and puzzles before. OpenAI’s ChatGPT (among others) has been totally crushed at chess by the computer opponent in a 1979 Atari game. A recent research paper from Apple found that models can struggle with other puzzles, like the Tower of Hanoi.

It has to do with the way LLMs work and fill in gaps in information. These models try to complete those gaps based on what happens in similar cases in their training data or other things they’ve seen in the past. With a sudoku, the question is one of logic. The AI might try to fill each gap in order, based on what seems like a reasonable answer, but to solve it properly, it instead has to look at the entire picture and find a logical order that changes from puzzle to puzzle.

Read more: 29 Ways You Can Make Gen AI Work for You, According to Our Experts

Chatbots are bad at chess for a similar reason. They find logical next moves but don’t necessarily think three, four or five moves ahead — the fundamental skill needed to play chess well. Chatbots also sometimes tend to move chess pieces in ways that don’t really follow the rules or put pieces in meaningless jeopardy.

You might expect LLMs to be able to solve sudoku because they’re computers and the puzzle consists of numbers, but the puzzles themselves are not really mathematical; they’re symbolic. «Sudoku is famous for being a puzzle with numbers that could be done with anything that is not numbers,» said Fabio Somenzi, a professor at CU and one of the research paper’s authors.

I used a sample prompt from the researchers’ paper and gave it to ChatGPT. The tool showed its work, and repeatedly told me it had the answer before showing a puzzle that didn’t work, then going back and correcting it. It was like the bot was turning in a presentation that kept getting last-second edits: This is the final answer. No, actually, never mind, this is the final answer. It got the answer eventually, through trial and error. But trial and error isn’t a practical way for a person to solve a sudoku in the newspaper. That’s way too much erasing and ruins the fun.

AI struggles to show its work

The Colorado researchers didn’t just want to see if the bots could solve puzzles. They asked for explanations of how the bots worked through them. Things did not go well.

Testing OpenAI’s o1-preview reasoning model, the researchers saw that the explanations — even for correctly solved puzzles — didn’t accurately explain or justify their moves and got basic terms wrong.

«One thing they’re good at is providing explanations that seem reasonable,» said Maria Pacheco, an assistant professor of computer science at CU. «They align to humans, so they learn to speak like we like it, but whether they’re faithful to what the actual steps need to be to solve the thing is where we’re struggling a little bit.»

Sometimes, the explanations were completely irrelevant. Since the paper’s work was finished, the researchers have continued to test new models released. Somenzi said that when he and Trivedi were running OpenAI’s o4 reasoning model through the same tests, at one point, it seemed to give up entirely.

«The next question that we asked, the answer was the weather forecast for Denver,» he said.

(Disclosure: Ziff Davis, CNET’s parent company, in April filed a lawsuit against OpenAI, alleging it infringed Ziff Davis copyrights in training and operating its AI systems.)

Explaining yourself is an important skill

When you solve a puzzle, you’re almost certainly able to walk someone else through your thinking. The fact that these LLMs failed so spectacularly at that basic job isn’t a trivial problem. With AI companies constantly talking about «AI agents» that can take actions on your behalf, being able to explain yourself is essential.

Consider the types of jobs being given to AI now, or planned for in the near future: driving, doing taxes, deciding business strategies and translating important documents. Imagine what would happen if you, a person, did one of those things and something went wrong.

«When humans have to put their face in front of their decisions, they better be able to explain what led to that decision,» Somenzi said.

It isn’t just a matter of getting a reasonable-sounding answer. It needs to be accurate. One day, an AI’s explanation of itself might have to hold up in court, but how can its testimony be taken seriously if it’s known to lie? You wouldn’t trust a person who failed to explain themselves, and you also wouldn’t trust someone you found was saying what you wanted to hear instead of the truth.

«Having an explanation is very close to manipulation if it is done for the wrong reason,» Trivedi said. «We have to be very careful with respect to the transparency of these explanations.»

Technologies

Today’s Wordle Hints, Answer and Help for Nov. 4, #1599

Here are hints and the answer for today’s Wordle for Nov. 4, No. 1,599.

Looking for the most recent Wordle answer? Click here for today’s Wordle hints, as well as our daily answers and hints for The New York Times Mini Crossword, Connections, Connections: Sports Edition and Strands puzzles.

Today’s Wordle puzzle begins with one of the least-used letters in the alphabet. (Check our full list ranking the letters by popularity.) If you need a new starter word, check out our list of which letters show up the most in English words. If you need hints and the answer, read on.

Today’s Wordle hints

Before we show you today’s Wordle answer, we’ll give you some hints. If you don’t want a spoiler, look away now.

Wordle hint No. 1: Repeats

Today’s Wordle answer has one repeated letter.

Wordle hint No. 2: Vowels

Today’s Wordle answer has two vowels, but one is the repeated letter, so you’ll see that one twice.

Wordle hint No. 3: First letter

Today’s Wordle answer begins with V.

Wordle hint No. 4: Last letter

Today’s Wordle answer ends with E.

Wordle hint No. 5: Meaning

Today’s Wordle answer can refer to the place where something happens, especially an organized event such as a concert, conference, or sports event.

TODAY’S WORDLE ANSWER

Today’s Wordle answer is VENUE.

Yesterday’s Wordle answer

Yesterday’s Wordle answer, Nov. 3, No. 1598 was AWOKE.

Recent Wordle answers

Oct. 30, No. 1594: LATHE

Oct. 31, No. 1595: ABHOR

Nov. 1, No. 1596: MOTEL

Nov. 2, No. 1597: RABID

Technologies

Why You Should Consider a Burner Phone for Your Holiday Travel This Year

If you’re traveling internationally, carrying a simple phone that doesn’t store personal information can be a smart move when entering the US.

Travel is challenging enough, and this year adds a new hurdle. US border agents are stepping up searches of travelers entering the country — even US citizens returning from overseas — and that extends to their personal devices. These searches can go beyond a quick look, giving agents the authority to copy or analyze a phone’s contents.

According to new figures from US Customs and Border Protection, nearly 15,000 device searches were carried out between April and June, with over 1,000 of them using advanced tools that copy or analyze what’s on a phone. The rising numbers raise questions about how much personal data travelers may be handing over without realizing it.

So what’s the solution? A burner phone. It’s the ultimate defense for keeping your personal data private when you travel, ensuring you stay connected without handing over your entire digital life at the border.

But the appeal goes beyond privacy. A stripped-down phone is also the perfect escape from the constant notifications and screen-time vortex of your primary device. Even celebrities such as Conan O’Brien have embraced simpler phones to cut through the noise. Whether you’re crossing a border or just trying to cross the street without distractions, a burner might be the smartest tech you own.

Read more: Best Prepaid Phone of 2025

Although carriers have offered prepaid phones since the ’90s, «burner phones» or «burners» became popular in the 2000s following the celebrated HBO series The Wire, where they helped characters avoid getting caught by the police. Although often portrayed in that light, burners aren’t only used by criminals; they’re also used anyone concerned with surveillance or privacy infringement.

What is a burner phone, and how does it work? Here’s everything you need to know about burners and how to get one.

Don’t miss any of our unbiased tech content and lab-based reviews. Add CNET as a preferred Google source.

What is a burner phone?

A burner phone is a cheap prepaid phone with no commitments. It comes with a set number of prepaid call minutes, text messages or data, and it’s designed to be disposed of after use.

Burners are contract-free, and you can grab them off the counter. They’re called burner phones because you can «burn» them (trash them) after use, and the phone can’t be traced back to you, which makes them appealing to criminals. Burner phones are typically used when you need a phone quickly, without intentions of long-term use.

Burners are different from getting a regular, contract-bound cellphone plan that requires your information to be on file.

Why should you use a burner phone?

Burner phones are an easy way to avoid cellphone contracts or spam that you get on your primary phone number. Burners aren’t linked to your identity, so you can avoid being tracked down or contacted.

You don’t have to dispose of a burner phone after use. You can add more minutes and continue using it. Burner phones can still function as regular phones, minus the hassle of a contract.

You can also get a burner phone as a secondary phone for a specific purpose, like having a spare phone number for two-factor authentication texts, for business, or to avoid roaming charges while traveling. Burner phones are often used by anyone concerned with privacy.

Burner phones, prepaid phones, smartphones and burner SIMs: What’s the difference?

Burner phones are cheap phones with simple designs that lack the bells and whistles of a smartphone. Because they’re designed to be disposable, you only get the essentials, as seen by the most common version, the flip phone.

All burner phones are prepaid phones, but not all prepaid phones are burners. What sets a burner apart is that you won’t have to give away any personal information to get one, and it won’t be traceable back to you. Again, a burner phone is cheap enough to be destroyed after use.

Prepaid smartphones are generally low-end models. You can use any unlocked smartphone with prepaid SIM cards, essentially making it a prepaid phone.

If you want a burner, you don’t necessarily have to buy a new phone. You can get a burner SIM and use it with an existing phone. Burner SIMs are prepaid SIMs you can get without a contract or giving away personal information.

Where can you buy a burner phone?

Burner phones are available at all major retail outlets, including Best Buy, Target and Walmart. They’re also often available at convenience stores like 7-Eleven, local supermarkets, gas stations and retail phone outlets like Cricket and Metro.

You can get a burner phone with cash, and it should cost between $10 and $50, although it may cost more if you get more minutes and data. If you’re getting a burner phone specifically to avoid having the phone traced back to you, it makes sense to pay with cash instead of a credit card.

If you just want a prepaid secondary phone, you can use a credit card. Just keep in mind that credit cards leave a trail that leads back to you.

There are also many apps that let you get secondary phone numbers, including Google Fi and the Burner app. However, these aren’t burners necessarily because the providers typically have at least some of your personal information.

If you’re just looking to get a solid prepaid phone without anonymity, check out our full guide for the best prepaid phone plans available. We also have a guide for the best cheap phone plans.

Technologies

Chrome Autofill Now Supports Passport, Driver’s License and Vehicle Info

Soon, you’ll never need to remember anything ever again.

Computer users are accustomed to web browsers autofilling everything from names and addresses to credit card numbers. Now, Google Chrome is adding new enhanced autofill options that allow users to automatically populate fields for passports, driver’s licenses, and their vehicle’s license plate or VIN, Google said in a blog post on Monday.

Desktop users must choose to turn on the feature, which is called enhanced autofill. Otherwise, it stays off. To turn it on, open Chrome, and at the top right of your browser, select more, then settings, then autofill and passwords. Finally, choose enhanced autofill and turn it in.

Google says Chrome now can «better understand complex forms and varied formatting requirements, improving accuracy across the web.» The company also says that enhanced autofill will be «private and secure.»

This enhanced autofill update is available in all languages, and more data options will be supported in the coming months.

A representative for Google said the company had no additional comment.

Don’t miss any of our unbiased tech content and lab-based reviews. Add CNET as a preferred Google source.

Chrome is a critical component in Google’s business. The web browser, currently the most popular in the world with a 73% market share, according to GlobalStats, provides the company with valuable user data that it uses to sell advertising. Advertising is how Google makes the majority of its revenues. New features help keep users loyal to Chrome, making it more difficult for them to switch to other browsers, including those from companies like Perplexity and OpenAI.