Do AI Browsers Exist For You - or To Give AI Companies Data? (fastcompany.com) 39
"It's been hard for me to understand why Atlas exists," writes MIT Technology Review. " Who is this browser for, exactly? Who is its customer? And the answer I have come to there is that Atlas is for OpenAI. The real customer, the true end user of Atlas, is not the person browsing websites, it is the company collecting data about what and how that person is browsing."
New York Magazine's "Intelligencer" column argues OpenAI wants ChatGPT in your browser because "That's where people who use computers, particularly for work, spend all their time, and through which vast quantities of valuable information flow in and out. Also, if you're a company hoping to train your models to replicate a bunch of white-collar work, millions of browser sessions would be a pretty valuable source of data."
Unfortunately, warns Fast Company, ChatGPT Atlas, Perplexity Comet, and other AI browses "include some major security, privacy, and usability trade-offs... Most of the time, I don't want to use them and am wary of doing so..." Worst of all, these browsers are security minefields. A web page that looks benign to humans can includehidden instructions for AI agents, tricking them into stealing info from other sites... "If you're signed into sensitive accounts like your bank or your email provider in your browser, simply summarizing a Reddit postcould result in an attacker being able to steal money or your private data,"Brave's security researchers wrotelast week.No one has figured out how to solve this problem.
If you can look past the security nightmares, the actual browsing features are substandard. Neither ChatGPT Atlas nor Perplexity Comet support vertical tabs — a must-have feature for me — and they have no tab search tool or way to look up recently-closed pages. Atlas also doesn't support saving sites as web apps, selecting multiple tabs (for instance, to close all at once with Cmd+W), or customizing the appearance. Compared to all the fancy new AI features, the web browsing part can feel like an afterthought. Regular web search can also be a hassle, even though you'll probably need it sometimes. When I typed "Sichuan Chili" into ChatGPT Atlas, it produced a lengthy description of the Chinese peppers, not the nearby restaurant whose website and number I was looking for.... Meanwhile, the standard AI annoyances still apply in the browser. Getting Perplexity to fill my grocery cart felt like a triumph, but on other occasions the AI has run into inexplicable walls and only ended up wasting more time.
There may be other costs to using these browsers as well. AI still has usage limits, and so all this eventually becomes a ploy to bump more people into paid tiers. Beyond that,Atlas is constantly analyzing the pages you visit to build a "memory" of who you are and what you're into. Do not be surprised if this translates to deeply targeted ads as OpenAI startslooking at ways to monetize free users. For now, I'm only using AI browsers in small doses when I think they can solve a specific problem.
Even then, I'm not going sign them into my email, bank accounts, or any other accounts for which a security breach would be catastrophic. It's too bad, because email and calendars are areas where AI agents could be truly useful, but the security risks are too great (andwell-documented).
The article notes that in August Vivaldi announced that "We're taking a stand, choosing humans over hype" with their browser: We will not use an LLM to add a chatbot, a summarization solution or a suggestion engine to fill up forms for you, until more rigorous ways to do those things are available. Vivaldi is the haven for people who still want to explore. We will continue building a browser for curious minds, power users, researchers, and anyone who values autonomy. If AI contributes to that goal without stealing intellectual property, compromising privacy or the open web, we will use it. If it turns people into passive consumers, we will not...
We're fighting for a better web.
New York Magazine's "Intelligencer" column argues OpenAI wants ChatGPT in your browser because "That's where people who use computers, particularly for work, spend all their time, and through which vast quantities of valuable information flow in and out. Also, if you're a company hoping to train your models to replicate a bunch of white-collar work, millions of browser sessions would be a pretty valuable source of data."
Unfortunately, warns Fast Company, ChatGPT Atlas, Perplexity Comet, and other AI browses "include some major security, privacy, and usability trade-offs... Most of the time, I don't want to use them and am wary of doing so..." Worst of all, these browsers are security minefields. A web page that looks benign to humans can includehidden instructions for AI agents, tricking them into stealing info from other sites... "If you're signed into sensitive accounts like your bank or your email provider in your browser, simply summarizing a Reddit postcould result in an attacker being able to steal money or your private data,"Brave's security researchers wrotelast week.No one has figured out how to solve this problem.
If you can look past the security nightmares, the actual browsing features are substandard. Neither ChatGPT Atlas nor Perplexity Comet support vertical tabs — a must-have feature for me — and they have no tab search tool or way to look up recently-closed pages. Atlas also doesn't support saving sites as web apps, selecting multiple tabs (for instance, to close all at once with Cmd+W), or customizing the appearance. Compared to all the fancy new AI features, the web browsing part can feel like an afterthought. Regular web search can also be a hassle, even though you'll probably need it sometimes. When I typed "Sichuan Chili" into ChatGPT Atlas, it produced a lengthy description of the Chinese peppers, not the nearby restaurant whose website and number I was looking for.... Meanwhile, the standard AI annoyances still apply in the browser. Getting Perplexity to fill my grocery cart felt like a triumph, but on other occasions the AI has run into inexplicable walls and only ended up wasting more time.
There may be other costs to using these browsers as well. AI still has usage limits, and so all this eventually becomes a ploy to bump more people into paid tiers. Beyond that,Atlas is constantly analyzing the pages you visit to build a "memory" of who you are and what you're into. Do not be surprised if this translates to deeply targeted ads as OpenAI startslooking at ways to monetize free users. For now, I'm only using AI browsers in small doses when I think they can solve a specific problem.
Even then, I'm not going sign them into my email, bank accounts, or any other accounts for which a security breach would be catastrophic. It's too bad, because email and calendars are areas where AI agents could be truly useful, but the security risks are too great (andwell-documented).
The article notes that in August Vivaldi announced that "We're taking a stand, choosing humans over hype" with their browser: We will not use an LLM to add a chatbot, a summarization solution or a suggestion engine to fill up forms for you, until more rigorous ways to do those things are available. Vivaldi is the haven for people who still want to explore. We will continue building a browser for curious minds, power users, researchers, and anyone who values autonomy. If AI contributes to that goal without stealing intellectual property, compromising privacy or the open web, we will use it. If it turns people into passive consumers, we will not...
We're fighting for a better web.
Give AI Companies Data? (Score:5, Insightful)
Re: (Score:1)
TBH I can't really conceive of why the question needed to be asked.
Like the Old Saying Goes . . . (Score:2)
Or with AI I guess your data is the product, and anything you might incidentally accomplish (learning, working, playing, earning, etc.) is a waste product, to be minimized accordingly.
There's probably car analogy that could express it better.
Re: (Score:2)
I don't mind giving Google data, but I don't give them everything. I mostly ask trivial questions related to e.g. cleaning, cooking, computers, math, chemistry, things I wouldn't mind even if someone finds out I asked those. If they can use that to improve their AI that is even better for me.
AI has been a great help for me, but I wouldn't pay for it.
It's supposed to be both (Score:4, Interesting)
If the data is worth something then the companies are supposed to offer you something compelling in value enough to you for you to give it to them, at least that's what a market capitalist should answer to that but that's a theoretical scenario where the consumer has the knowledge and market power to make an informed choice including most importantly the ability to walk away from a purchase.
It is up to the individual to determine the value the product gives but I would say it's effectively impossible for that user to make that choice since it's unreasonable for the user to know every bit of data collected, how it's stored, where it gets shared or maintain control over it.
It seems in the data market that the data itself has become so valuable and the market for it is so opaque that there is effectively a market failure so that's where regulation is going to have to step in. This obviously isn't just an AI thing but AI is just the next step in industrializing the market for user data.
Personally, much like cryptocurrency I see the personal user data market and all it entails as offering little so value to the average person in terms of practical utility that we can outright ban it or regulate it so heavily to cause a large retraction in its size. They both exist in a loop simply to support itself because there are so many dollars in it now.
Re: It's supposed to be both (Score:2)
Re: (Score:2)
"...at least that's what a market capitalist should answer..."
What's a market capitalist? If you mean capitalist, you are clearly wrong. If you mean believer in free markets, I believe you are also wrong. What you are describing is ethical behavior, capitalism is not about ethical behavior and free markets are supposed to apply pressure to capitalism to cause ethical behaviors (but not require them).
"...most importantly the ability to walk away from a purchase."
That requires free markets. We don't have
Re: (Score:2)
What you are describing is ethical behavior, capitalism is not about ethical behavior and free markets are supposed to apply pressure to capitalism to cause ethical behaviors (but not require them).
Nothing about what I described requires any form of ethics, in fact it is distinctly outside of ethics, it's simple supply/demand.
Also which is it, capitalism nor free markets don't apply any ethical pressure only pressure is what leads to profits. Now this *should* line up with ethics and sometimes it does naturally but sometimes it doesn't because reality exists and you have things like market failures, perverse incentives, rent seeking behavior, etc etc. That's usually what we mean by regulations appl
Re: (Score:2)
"Simple supply and demand" requires ethics, or it doesn't exist.
If your browser can do anything you can, it will (Score:5, Insightful)
Irritating! (Score:4, Informative)
I am increasingly bombarded by AI generated text. It seems that browsers, search engines, customer "service" and many app have incorporated AI. Even my Tesla now has the Nazi Grok AI as default.
Besides the obvious threat of monitoring my thoughts and behavior, I find all of these AI "services" very irritating. They serve up long winded rambling blocks of "information" which is usually at least partially wrong or irrelevant.
I really hate all of these AI efforts.
Did you really have to ask? (Score:2)
The title of TFS is so clearly and perfectly a "rhetorical question", that it could easily be used as an example of same in an official definition of the term.
I tried Comet (Score:3)
I couldn't find any use for it
All of the examples they listed seemed silly and useless
Re: (Score:2)
This version actually is useful.
https://www.cometcleaner.com/ [cometcleaner.com]
AI companies are... (Score:3)
... content and privacy rapists. Period. Draw the appropriate conclusions.
Anyone who needs to ask that question (Score:2)
needs a keeper to run their life for them.
Nobody could possibly be so disconnected from reality as to not know the answer to that, and still be able to walk and breathe at the same time without help.
Re: Anyone who needs to ask that question (Score:4, Insightful)
Re: (Score:2)
I know you are, but what am I?
This headline should be a sentence. (Score:2)
There is no reason to question that's what it's for.
Two usage scenarios come to mind. (Score:2)
First, AI companies want to know what problems people are using their computers to solve, so that they can train their LLMs to solve those problems, and then sell subscriptions to the resulting services.
Second, the AI companies know that high quality content is getting harder to find (because their own products are generating so much garbage). Watching user behavior may help them identify the pages that users find helpful, and pages that users find worthless. They will presumably use that information to gui
Re: (Score:2)
If they want high quality content, they should set up a user moderated site (like Slashdot) covering general news of interest. Perhaps several covering different areas. And have part of the TOS that they can used the data to train an AI. That kind of site is provably cheap to operate, and can even have a few ads. And the "user moderation" will provide training feedback to the AI. (I don't think reddit is as good a model, even though it's already being used that way.)
Re: Two usage scenarios come to mind. (Score:2)
The tricky thing is distinguishing between truly user generated content and AI slop.
Not that a browser completely solves that problem but at least it raises the bar a little. When the input gestures are consistent with typing then that seems like a reason to trust it more than something that was pasted in.
Still spoofable, but there's much less incentive for users to spoof from that direction. At least, until it becomes clear that there's an ROI for impersonating a human user of openai's browser. Which, even
Re: (Score:2)
If they want high quality content, they should set up a user moderated site (like Slashdot) covering general news of interest. Perhaps several covering different areas. And have part of the TOS that they can used the data to train an AI. That kind of site is provably cheap to operate, and can even have a few ads. And the "user moderation" will provide training feedback to the AI. (I don't think reddit is as good a model, even though it's already being used that way.)
Isn't this what Reddit did?
Re: (Score:2)
I consider the Slashdot model to be a LOT better than Reddit. (I rarely even look at Reddit.)
Re: (Score:2)
I consider the Slashdot model to be a LOT better than Reddit. (I rarely even look at Reddit.)
I prefer Slashdot too. Moderation points are capped, which seems to reduce the karma-whoring. I wish Slashdot would bring back true anonymous posting though.
Reddit has forums, whereas Slashdot is "just" a news aggregator. This means that Slashdot has a regular turnover of threads, and old ones are locked. Depending on what you want, this may be better or worse. Different jobs, different tools.
I rarely look at Reddit either. It's too tiresome to edit the URL every time to get the "old" format.
Apply the Basic Test (Score:4, Interesting)
Did you pay money for it? If the answer is no, then you are not the customer; you are the product! So applying that, this browser is designed for you to train the AI.
This works for everything online. Though I suppose there are some additional grades. Did you pay ENOUGH to sustain your usage? If the answer is no then you are partially the product. Companies that offer "free" things are evil little spies. Companies that offer subscription services that cost money are potentially not stealing your data or tricking you into doing things for them.
FUD (Score:2)
"...if you're a company hoping to train your models to replicate a bunch of white-collar work, millions of browser sessions would be a pretty valuable source of data."
And it doesn't need to be an AI browser. In fact, it can be browser-agnostic and the ability to track those sessions has existed for a long time.
"A web page that looks benign to humans can include hidden instructions for AI agents..."
But that would require that the web page itself be submitted to the AI, the web page is rendered by the render
is that a rhetorical question? (Score:2)
AI belongs to the platform holders (Score:2)
You need to be able to tell the difference between slop and real human generated content.
You can easily do that with advanced user metrics. It's genuinely difficult to fake the erratic behavior of human beings with things like keyboard typing and mouse movement among other things.
But to get that low levels you need to be the platform holder.
So it's no surprise big AI companies are trying to m
Why is that even a question? (Score:2)
All the AI companies care for is make tons of money and make their owners rich. The somewhat crappy product they push with lies, lies and more lies make that abundantly clear. On top of that, they are currently in the process of creating a really spectacular economic crash that will hur a lot of people. Actually good people do not do these things.
Re: (Score:2)
Neither (Score:2)
They exist to give the AI companies customers. The user data is mostly worthless. Crawling is faster with crawlers, they won't need to sneak a secret crawler into the browser. User inputs are low quality and the AI sites get them anyway. But having the user use the own AI instead of the rival's is worth a lot. Google builds a whole own browser to make sure the user searches with Google, of course the AI companies like to direct users to their AI site using a cheap chrome fork.
the Comet browser (Score:2)
I was using the Perplexity app at times but switched over to their Comet browser. Its basically the same thing but with tabs. I use it to answer technical questions where it would take me a lot of time to sift through ordinary search engine results. It works pretty well for that and provides links to relevant articles if you want to dig deeper. It isn't my every day browser by any means but I've found that it does give good quick answers for certain kinds of queries. Yesterday I used it to investigate a bra
This is the purpose of AI agents (Score:2)
No different than "classic" browsers (Score:2)
Chrome, Edge, Safari, Firefox, they all send usage data to the companies that made the browsers.