Connect with us

Technologies

AI as Lawyer: It’s Starting as a Stunt, but There’s a Real Need

People already have a hard enough time getting help from lawyers. Advocates say AI could change that.

Next month, AI will enter the courtroom, and the US legal system may never be the same.

An artificial intelligence chatbot, technology programmed to respond to questions and hold a conversation, is expected to advise two individuals fighting speeding tickets in courtrooms in undisclosed cities. The two will wear a wireless headphone, which will relay what the judge says to the chatbot being run by DoNotPay, a company that typically helps people fight traffic tickets through the mail. The headphone will then play the chatbot’s suggested responses to the judge’s questions, which the individuals can then choose to repeat in court.

It’s a stunt. But it also has the potential to change how people interact with the law, and to bring many more changes over time. DoNotPay CEO Josh Browder says expensive legal fees have historically kept people from hiring traditional lawyers to fight for them in traffic court, which typically involves fines that can reach into the hundreds of dollars.

So, his team wondered whether an AI chatbot, trained to understand and argue the law, could intervene.

«Most people can’t afford legal representation,» Browder said in an interview. Using the AI in a real court situation «will be a proof of concept for courts to allow technology in the courtroom.»

Regardless of whether Browder is successful — he says he will be — his company’s actions mark the first of what are likely to be many more efforts to bring AI further into our daily lives.

Modern life is already filled with the technology. Some people wake up to a song chosen by AI-powered alarms. Their news feed is often curated by a computer program, too, one that’s taught to pick items they’ll find most interesting or that they’ll be most likely to comment on and share via social media. AI chooses what photos to show us on our phones, it asks us if it should add a meeting to our calendars based on emails we receive, and it reminds us to text a birthday greeting to our loved ones.

But advocates say AI’s ability to sort information, spot patterns and quickly pull up data means that in a short time, it could become a «copilot» for our daily lives. Already, coders on Microsoft-owned GitHub are using AI to help them create apps and solve technical problems. Social media managers are relying on AI to help determine the best time to post a new item. Even we here at CNET are experimenting with whether AI can help write explainer-type stories about the ever-changing world of finance.

So, it can seem like only a matter of time before AI finds its way into research-heavy industries like the law as well. And considering that 80% of low-income Americans don’t have access to legal help, while 40% to 60% of the middle class still struggle to get such assistance, there’s clearly demand. AI could help meet that need, but lawyers shouldn’t feel like new technology is going to take business away from them, says Andrew Perlman, dean of the law school at Suffolk University. It’s simply a matter of scale.

«There is no way that the legal profession is going to be able to deliver all of the legal services that people need,» Perlman said.

Turning to AI

DoNotPay began its latest AI experiment back in 2021 when businesses were given early access to GPT-3, the same AI tool used by the startup OpenAI to create ChatGPT, which went viral for its ability to answer questions, write essays and even create new computer programs. In December, Browder pitched his idea via a tweet: have someone wear an Apple AirPod into traffic court so that the AI could hear what’s happening through the microphone and feed responses through the earbud.

Aside from people jeering him for the stunt, Browder knew he’d have other challenges. Many states and districts limit legal advisors to those who are licensed to practice law, a clear hurdle that UC Irvine School of Law professor Emily Taylor Poppe said may cause trouble for DoNotPay’s AI.

«Because the AI would be providing information in real time, and because it would involve applying relevant law to specific facts, it is hard to see how it could avoid being seen as the provision of legal advice,» Poppe said. Essentially, the AI would be legally considered a lawyer acting without a law license.

AI tools raise privacy concerns too. The computer program technically needs to record audio to interpret what it hears, a move that’s not allowed in many courts. Lawyers are also expected to follow ethics rules that forbid them from sharing confidential information about clients. Can a chatbot, designed to share information, follow the same protocols?

Perlman says many of these concerns can be answered if these tools are created with care. If successful, he argues, these technologies could also help with the mountains of paperwork lawyers encounter on a daily basis.

Ultimately, he argues, chatbots may turn out to be as helpful as Google and other research tools are today, saving lawyers from having to physically wade through law libraries to find information stored on bookshelves.

«Lawyers trying to deliver legal services without technology are going to be inadequate and insufficient to meeting the public’s legalities,» Perlman said. Ultimately, he believes, AI can do more good than harm.

The two cases DoNotPay participates in will likely impact much of that conversation. Browder declined to say where the proceedings will take place, citing safety concerns.

Neither DoNotPay nor the defendants plan to inform the judges or anyone in court that an AI is being used or that audio is being recorded, a fact that raises ethics concerns. This in itself resulted in pushback on Twitter when Browder asked for traffic ticket volunteers in December. But Browder says the courts that DoNotPay chose are likely to be more lenient if they find out.

The future of law

After these traffic ticket fights, DoNotPay plans to create a video presentation designed to advocate in favor of the technology, ultimately with the goal of changing law and policy to allow AI in courtrooms.

States and legal organizations, meanwhile, are already debating these questions. In 2020, a California task force dedicated to exploring ways to expand access to legal services recommended allowing select unlicensed practitioners to represent clients, among other reforms. The American Bar Association told judges using AI tools to be mindful of biases instilled in the tools themselves. UNESCO, the international organization dedicated to preserving culture, has a free online course covering the basics of what AI can offer legal systems.

For his part, Browder says AI chatbots will become so popular in the next couple of years that the courts will have no choice but to allow them anyway. Perhaps AI tools will have a seat at the table, rather than having to whisper in our ears.

«Six months ago, you couldn’t even imagine that an AI could respond in these detailed ways,» Browder said. «No one has imagined, in any law, what this could be like in real life.»

Related Topics:

Technologies

Turns Out Perplexity Might Be the Sleeper Feature on Samsung’s Galaxy S26

Having Perplexity’s AI and models on devices from the world’s biggest phone-maker puts the company under a brighter light.

There were plenty of references to AI at today’s Galaxy Unpacked event. But Samsung isn’t alone; nearly every major smartphone launch in recent years has included new AI features or partnerships with AI companies.

Samsung launched its latest iteration of Galaxy AI, debuting it alongside Galaxy S26 phones. This follows weekend news that the company plans to integrate Perplexity’s AI agent — and even support a «Hey Plex» wake word — on its new phones. But the partnership appears to go beyond simply giving Samsung users another AI option.

Since late 2023, phone-makers have been leapfrogging one another to add generative AI features and integrate AI agents. Nearly every new Android phone supports Google’s Gemini assistant. Apple’s iPhones integrate OpenAI’s ChatGPT into the phone’s Visual Intelligence feature and its Siri overhaul will incorporate Google’s Gemini AI models.

While Perplexity has partnered with phone-makers such as Motorola to preload its app — and has been integrated into devices for Deutsche Telekom — having its AI and models built directly into phones from the world’s largest manufacturer puts the company on a much bigger stage. It marks a shift toward AI agents being just another tool people choose to use, much like a phone app.

«The first step toward an agentic mobile ecosystem is the user getting to choose whatever agent they want,» Dmitry Shevelenko, Perplexity’s chief business officer, told CNET. «I think this is where Samsung is taking a big, big leap forward.»

Perplexity’s Sonar API powers aspects of Samsung’s Galaxy AI ecosystem. Shevelenko said that the company’s engineers worked closely with Samsung’s team to revamp its Bixby assistant at the framework level, getting deep system access. He noted that it’s the first time a third-party AI company has achieved parity on a major mobile OS. The Galaxy S26 phones that Samsung announced support the new «Hey Plex» wake word, putting Perplexity shoulder-to-shoulder with Google’s Gemini AI assistant, which is integrated into Android on Samsung devices.

«What’s unique is the only other company that has it is Google, right?» said Shevelenko. «It’s a real paradigm shift for Samsung to be going into a multi-AI direction, where they are giving their users choice. And I think they see this as a strategic differentiator.»

Samsung’s inclusion of Perplexity touches many of the company’s own apps including Calendar, Clock, Gallery, Notes and Reminders. The benefit of structuring Perplexity’s AI deeply into Samsung’s software is that people can have a lighter interaction with their phones. As opposed to unlocking their device, navigating the home page, opening the app and entering a query, people will be able to simply press a button, say, «Hey Plex,» starting their search within seconds.

But the integration of Perplexity isn’t limited to Bixby. Shevelenko said Samsung’s browser, aptly named Internet, includes agentic browsing using Perplexity’s Comet technology as well.

Such a significant moment for Perplexity naturally draws parallels to Apple and its partnership with OpenAI, which has partnered with former Apple designer Jony Ive for its own hardware efforts. When I asked Shevelenko about the possibility of Perplexity making its own phone or hardware, he responded emphatically, «No.»

«We are laser-focused on working with all the best OEMs,» he said. «Our thing we’re world-class at is building accurate AI that is easy to use and delightful to use and growing that curiosity.»

And while we wait for Samsung to announce new phones, it’ll be interesting to see how Galaxy phone owners use the phone’s AI agents. Soon, people could say, «Hey Google» into their Samsung devices to prompt Gemini, or «Hey Plex» to trigger a query with Perplexity. And options are usually a good thing.

Samsung did not immediately respond to a request for comment.

Continue Reading

Technologies

ADT Acquires AI Company for Sensing People and Activity in Your Home

ADT’s acquisition of Origin AI brings presence-sensing technology under the home security company’s umbrella.

ADT on Tuesday announced an interesting new acquisition for anyone looking to the future of home security — and it’s no surprise AI is a part of the story. In a $170 million deal, ADT has purchased Origin AI, which specializes in people detection in spaces like the inside of your home, something the security company is calling AI-sensing technology.

ADT has not disclosed specific plans for AI technology, but this comes at a time when concerns about corporate surveillance by companies like Ring and Flock have reached a fever pitch.

«ADT has been testing and evaluating Origin’s technology pre-acquisition,» ADT Chief Business Officer Omar Kahn told me. «In 2026, the focus is on integrating the technology into ADT’s platform, with commercialization expected to begin in 2027.»

Presence sensing doesn’t sound like the chatty, summary-creating large language models we consider AI these days, nor the person and car recognition features companies like Flock use. It’s a system that analyzes home Wi-Fi frequencies for disruptions. The AI is trained in pattern recognition to identify which disruptions indicate that humans are at home (ignoring pets) and what they may be doing.

The technology has cropped up in many spots over the past couple of years. I’ve seen it before with aging-in-place technology and Philips Hue’s newest smart bulbs, but most recently with Aqara’s sensor at CES 2026, which can detect when multiple people are congregating, standing, sitting or lying down. 

How does presence sensing affect people’s privacy?

It’s not clear how ADT will use Origin’s presence sensing in its home security systems, though the company did mention smart automation, personalization and reducing false alarms. In one example, it could automatically adjust an ADT-supported thermostat when multiple people are detected moving around a house. But that also raises privacy questions.

Presence sensing, like Origin’s tech, has certain privacy benefits. It doesn’t use cameras to film anyone or save video recordings of people, and it doesn’t create identity profiles based on someone’s face or other data. It can’t tell who is in a house, only where they are and how/when they are moving around (or not moving).

That allows for capabilities such as notifying a nursing home that a resident hasn’t gotten out of bed when they usually do, without invasive investigation. But the technology also raises privacy concerns: A company could know when people in their own home are in bed, watching TV, or sitting to eat dinner, even if it can’t identify them by name.

ADT calls features like these home awareness, but also mentions municipal compliance and coordination with first responders. That could mean giving firefighters information on how many people are in a burning building. But there are concerns. Recent news reports indicate that some local law enforcement agencies have shared information with US Immigration and Customs Enforcement for use in home and apartment raids, raising the possibility that the technology could be applied in similar contexts.

The technology’s implications may ultimately hinge on how ADT chooses to implement and regulate it. Until those details are clearer, its promise and its risks remain closely intertwined.

Continue Reading

Technologies

New York Times Debuts the Midi Crossword, Its In-Between Puzzle

Is the Mini Crossword too easy, but the original one just too time-consuming? Here’s your new puzzle.

The daily New York Times Mini Crossword can be solved in a minute or so, while the newspaper’s iconic original crossword puzzle might take hours. Now, puzzlers who want an in-between diversion can try a new puzzle from the Times, introduced this week — the Midi Crossword puzzle. (And CNET readers can get daily answers for five Times puzzles — Wordle, Connections, Strands, Connections: Sports Edition and the Mini Crossword.)

New York Times Games subscribers can play the Midi in the New York Times Games app for iOS and Android devices, or on mobile or desktop web. It’s online-only, not in the print newspaper. 

«We’re really leaning into the digital-first nature of the puzzle,» NYT Games Puzzle Editor Ian Livengood said in a Times article about the new puzzle. «About once a week, the puzzle will have a visual effect — an extra flourish when you start or after you solve. This could be a cool animation or colorful shading.»

As the name «Midi» suggests, this is a mid-sized crossword puzzle. Where the Mini Crossword usually only has 5 Across and 5 Down clues, the Midi is usually a 9-by-9 puzzle, sometimes as long as 11-by-11.

«If you feel like the Mini is not enough but the Daily is too much, this will be the perfect puzzle for you,» Livengood said.

Each Midi Crossword has a theme that hints at the topics of the clues and answers. Unlike the other puzzles, Livengood says the Midi might occasionally have two-letter words and repeating answers.

I tried the Midi Crossword

I tried Wednesday’s Midi Crossword and solved it in just over 3 minutes. That’s much longer than I spend on the Mini Crossword, but much faster than the original New York Times crossword puzzle takes me. 

I thought most of the clues were pretty simple, and the few tricky ones filled themselves in once I moved from Across to Down.

If you’re a New York Times Games subscriber, this is a nice addition to your daily puzzle stable. It tests your mind a bit more than the Mini, but you can also solve it while watching TV or waiting for someone to text you back.

Continue Reading

Trending