Is AI the future of web browsing?
If you don't remember, no one will blame you. Web browsers have remained fundamentally unchanged for decades: You open an app, such as Chrome, Safari or Firefox, and type a website into the address bar. Many of us settled on one and fell into what I call 'browser inertia,' never bothering to see if there's anything better.
Yet a web browser is important because so much of what we do on computers takes place inside one, including word processing, chatting on Slack and managing calendars and email.
That's why I felt excited when I recently tried Dia, a new kind of web browser from the Browser Co. of New York, a startup. The app is powered by generative artificial intelligence, the technology driving popular chatbots like ChatGPT and Google's Gemini, to answer our questions. Dia illuminates how a web browser can do much more than load websites – and even help us learn and save time.
I tested Dia for a week and found myself browsing the web in new ways. In seconds, the browser provided a written recap of a 20-minute video without my watching its entirety. While scanning a breaking news article, the browser generated a list of other relevant articles for a deeper understanding. I even wrote to the browser's built-in chatbot for help proofreading a paragraph of text.
Dia is on the cusp of an emerging era of AI-powered internet navigators that could persuade people to try something new. This week, Perplexity, a startup that makes a search engine, announced an AI web browser called Comet, and some news outlets have reported that OpenAI, the company behind ChatGPT, also plans to release a browser this year. OpenAI declined to comment. (The New York Times has sued OpenAI and its partner, Microsoft, claiming copyright infringement of news content related to AI systems. The two companies have denied the suit's claims.)
Tech behemoths like Google and Apple have added lightweight AI features into their existing browsers, Chrome and Safari, including tools for proofreading text and automatically summarizing articles.
Dia, which has not yet been publicly released, is available as a free app for Mac computers on an invitation-only basis.
What does this all mean for the future of the web? Here's what you need to know.
What is an AI browser, and what does it do?
Like other web browsers, Dia is an app you open to load webpages. What's unique is the way the browser seamlessly integrates an AI chatbot to help – without leaving the webpage.
Hitting a shortcut (command+E) in Dia opens a small window that runs parallel to the webpage. Here, you can type questions related to the content you are reading or the video you are watching, and a chatbot will respond.
For example:
– While writing this column on the Google Docs website, I asked the chatbot if I used 'on the cusp' correctly, and it confirmed that I did.
– While reading a news article about the Texas floods, I asked the browser's chatbot to tell me more about how the crisis unfolded. The bot generated a summary about the history of Texas' public safety infrastructure and included a list of relevant articles.
– While watching a 22-minute YouTube video about car jump starters, I asked the chatbot to tell me which tools were best. Dia immediately pulled from the video's transcript to produce a summary of the top contenders, sparing me the need to watch the entire thing.
In contrast, chatbots like ChatGPT, Gemini and Claude require opening a separate tab or app and pasting in content for the chatbot to evaluate and answer questions, a process that has always busted my workflow.
How does it work?
AI chatbots like ChatGPT, Gemini and Claude generate responses using large language models, systems that use complex statistics to guess which words belong together. Each chatbot's model has its strengths and weaknesses.
The Browser Co. of New York said it had teamed up with multiple companies to use their AI models, including the ones behind Gemini, ChatGPT and Claude. When users type a question, the Dia browser analyzes it and pulls answers from whichever AI model is best suited for answering.
For instance, Anthropic's AI model, Claude Sonnet, specializes in computer programming. So if you have questions about something you are coding, the browser will pull an answer from that model. If you have questions about writing, the Dia browser may generate an answer with the model that OpenAI uses for ChatGPT, which is well known for handling language.
What I appreciate about this design is that you, the user, don't need to know or think about which chatbot to use. That makes generative AI more accessible to the mainstream.
'You should just be able to say, 'Hey, I'm looking at this thing, I've got a question about it,'' said Josh Miller, the CEO of the Browser Co., which was founded in 2020 and has raised over $100 million. 'We should be able to answer it for you and do work on your behalf.'
But aren't there imperfections?
While Dia proved helpful in most of my tests, it was, like all generative AI tools, sometimes incorrect.
While I was browsing Wirecutter, a New York Times publication that reviews products, I asked the chatbot if there were any deals on the site for water filters. The chatbot said no, even as I read about a water filtration system that was on sale.
Miller said that because the browser drew answers from various AI models, its responses were subject to the same mistakes as their respective chatbots. Those occasionally get facts wrong and even make things up, a phenomenon known as 'hallucination.'
More often than not, however, I found Dia to be more accurate and helpful than a stand-alone chatbot. Still, I double-checked answers by clicking on any links Dia's bot was citing, like the articles about the recent floods in Texas.
What about privacy?
Asking AI to help with a webpage you're looking at means that data may be shared with whatever AI model is being used to answer the question, which raises privacy concerns.
The Browser Co. said that only the necessary data related to your requests was shared with its partners providing AI models, and that those partners were under contract to dispose of your data.
Privacy experts have long warned not to share any sensitive information, like a document containing trade secrets, with an AI chatbot since a rogue employee could gain access to the data.
So I recommend asking Dia's chatbot for help only with innocuous browsing activities like parsing a YouTube video. But when browsing something you wouldn't want others to know about, like a health condition, refrain from using the AI.
This exchange – potentially giving up some privacy to get help from AI – may be the new social contract going forward.
How much will this cost?
Dia is free, but AI models have generally been very expensive for companies to operate. Consumers who rely on Dia's AI browser will eventually have to pay.
Miller said that in the coming weeks, Dia would introduce subscriptions costing US$5 a month to hundreds of dollars a month, depending on how frequently a user prods its AI bot with questions. The browser will remain free for those who use the AI tool only a few times a week.
So whether an AI browser will be your next web browser will depend largely on how much you want to use, and pay, for these services. So far, only 3% of the people who use AI every day are paid users, according to a survey by Menlo Ventures, a venture capital firm.
That number could grow, of course, if generative AI becomes a more useful tool that we naturally use in everyday life. I suspect the humble web browser will open that path forward. – ©2025 The New York Times Company
This article originally appeared in The New York Times.
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
The Star
4 hours ago
- The Star
Alphabet hit with EU antitrust complaint by six digital rights groups
FILE PHOTO: The new Google logo is seen in this illustration taken May 13, 2025. REUTERS/Dado Ruvic/Illustration/File Photo BRUSSELS (Reuters) -Alphabet was targeted with an EU antitrust complaint from six human and digital rights groups on Thursday which urged EU regulators to investigate whether the tech giant complies with legislation requiring it to make it easier for users to uninstall software apps. The EU's Digital Markets Act (DMA), which took effect two years ago, sets out a list of dos and don'ts for seven Big Tech companies including Alphabet unit Google, in an attempt to curb their power and give rivals more room and users more choice. British human rights organisation ARTICLE 19, European Digital Rights (EDRi), Free Software Foundation Europe (FSFE), Gesellschaft fur Freiheitsrechte (GFF), Homo Digitalis, and said Alphabet allegedly has not complied with the DMA. They said the alleged violation concerned a DMA requirement that gatekeepers or companies which provide a core platform service to business users, shall technically enable users to easily uninstall software applications on the gatekeeper's operating system. "Alphabet has designed its Core Platform Service Android in a way to hide from end users the possibility to disable its own pre-installed gatekeeper apps," they said in their complaint. "What is more, Alphabet goes to great length to scare away end users who have found that possibility against all odds of actually disabling Google's pre-installed apps," they said. The group called on the European Commission to investigate the issue. Alphabet refuted the allegations. "It is easy to uninstall apps on Android devices, so this complaint does not represent a genuine user concern. Other regulators, including the CMA, have previously dismissed this complaint," a Google spokesperson said. The CMA is the British competition watchdog. The Commission confirmed receipt of the complaint, saying it is currently assessing it under its standard procedures. (Reporting by Foo Yun Chee; Editing by Susan Fenton)
The Star
12 hours ago
- The Star
AI is slowing down (not speeding up) software developers, study finds
AI tools have widely been adopted by programmers in efforts to speed up production, however new research is casting doubt on the assumption that it generally helps developers work faster. — Photo: Sebastian Gollnow/dpa LOS ANGELES: Open-source developers who lean on artificial intelligence tools are less productive and efficient than those who do not use the much-hyped technology, according to researchers investigating AI assistance in software development. Using AI means the work takes 19% longer to finish than without, said a team at METR (Model Evaluation and Threat Research) after a randomized control trial involving 16 developers who undertook a series of typical tasks their job entails, such as bug fixes. AI tools have widely been adopted by programmers in efforts to speed up production, and in May ChatGPT maker OpenAI released a dedicated software engineering agent called Codex for software developers. However: "AI makes them slower," the Californian non-profit METR said, deeming their outcome surprising - as did the developers, who said in advance that they expected using AI to make them around 25% faster at getting through their tasks. Not only was the "significant slowdown" contrary to developer beliefs and experts' forecasts, the "gap between perception and reality" was "striking," according to the researchers. Perhaps the most striking finding of all was that despite AI throwing sand in their gears as they worked, the developers nonetheless believed it "had sped them up by 20%," according to METR's team. "They were mistaken about AI's impact on their productivity," the METR team said of the participants. But the researchers by no means wanted to rule out AI as a helpful tool for coders and said it could still be useful for less experienced developers or those "working in an unfamiliar codebase." The findings echo research published by Stanford University earlier this year showing AI as largely hindering workers of experience, while it boosts those with less skill. The effects of AI on productivity "vary significantly," the researchers said. – dpa
The Star
13 hours ago
- The Star
Google AI system wins gold medal in International Mathematical Olympiad
SAN FRANCISCO: An artificial intelligence system built by Google DeepMind, the tech giant's primary AI lab, has achieved 'gold medal' status in the annual International Mathematical Olympiad, a premier math competition for high school students. It was the first time that a machine – which solved five of the six problems at the 2025 competition, held in Australia this month – reached that level of success, Google said in a blog post Monday. The news is another sign that leading companies are continuing to improve their AI systems in areas such as math, science and computer coding. This kind of technology could accelerate the research of mathematicians and scientists and streamline the work of experienced computer programmers. Two days before Google revealed its feat, an OpenAI researcher said in a social media post that the startup had built technology that achieved a similar score on this year's questions, although it did not officially enter the competition. Both systems were chatbots that received and responded to the questions much like humans. Other AI systems have participated in the International Mathematical Olympiad, or IMO, but they could answer questions only after human experts translated them into a computer programming language built for solving math problems. 'We solved these problems fully in natural language,' Thang Luong, a senior staff research scientist at Google DeepMind, said in an interview. 'That means there was no human intervention – at all.' After OpenAI started the AI boom with the release of ChatGPT in late 2022, the leading chatbots could answer questions, write poetry, summarise news articles, even write a little computer code. But they often struggled with math. Over the past two years, companies such as Google and OpenAI have built AI systems better suited to mathematics, including complex problems that the average person cannot solve. Last year, Google DeepMind unveiled two systems that were designed for math: AlphaGeometry and AlphaProof. Competing in the IMO, these systems achieved 'silver medal' performance, solving four of the competition's six problems. It was the first time a machine reached silver-medal status. Other companies, including a startup called Harmonic, have built similar systems. But systems such as AlphaProof and Harmonic are not chatbots. They can answer questions only after mathematicians translate the questions into Lean, a computer programming language designed for solving math problems. This year, Google entered the IMO with a chatbot that could read and respond to questions in English. This system is not yet available to the public. Called Gemini Deep Think, the technology is what scientists call a 'reasoning' system. This kind of system is designed to reason through tasks involving math, science and computer programming. Unlike previous chatbots, this technology can spend time thinking through complex problems before settling on an answer. Other companies, including OpenAI, Anthropic and China's DeepSeek, offer similar technologies. Like other chatbots, a reasoning system initially learns its skills by analysing enormous amounts of text culled from across the internet. Then it learns additional behaviour through extensive trial and error in a process called reinforcement learning. A reasoning system can be expensive, because it spends additional time thinking about a response. Google said Deep Think had spent the same amount of time with the IMO as human participants did: 4 1/2 hours. But the company declined to say how much money, processing power or electricity had been used to complete the test. In December, an OpenAI system surpassed human performance on a closely watched reasoning test called ARC-AGI. But the company ran afoul of competition rules because it spent nearly US$1.5mil (RM6.3mil) in electricity and computing costs to complete the test, according to pricing estimates. – ©2025 The New York Times Company This article originally appeared in The New York Times.
