惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

博客园 - 三生石上(FineUI控件)
月光博客
月光博客
S
SegmentFault 最新的问题
有赞技术团队
有赞技术团队
Stack Overflow Blog
Stack Overflow Blog
Engineering at Meta
Engineering at Meta
T
The Blog of Author Tim Ferriss
The GitHub Blog
The GitHub Blog
小众软件
小众软件
Hugging Face - Blog
Hugging Face - Blog
IT之家
IT之家
宝玉的分享
宝玉的分享
A
About on SuperTechFans
Vercel News
Vercel News
P
Proofpoint News Feed
freeCodeCamp Programming Tutorials: Python, JavaScript, Git & More
博客园 - 【当耐特】
钛媒体:引领未来商业与生活新知
钛媒体:引领未来商业与生活新知
V
Visual Studio Blog
Jina AI
Jina AI
Y
Y Combinator Blog
T
Tailwind CSS Blog
奇客Solidot–传递最新科技情报
奇客Solidot–传递最新科技情报
Last Week in AI
Last Week in AI

Lauren Weinstein's Blog

Big Tech’s Catastrophic Layoffs Frenzy – Lauren Weinstein's Blog The Rise of AI Slop on Google’s YouTube – Lauren Weinstein's Blog Why Google Search AI Overviews Are a “Misinformation Machine” – Lauren Weinstein's Blog Why Apple’s New “MacBook Neo” Has Stunned the Tech World – Lauren Weinstein's Blog When Data Centers Destroy Communities – Lauren Weinstein's Blog Why Proposed “Blocking Technologies” for 3D Printers Are a Terrible Idea – Lauren Weinstein's Blog Cheerful, Cooperative, and Usually, Wrong. – Lauren Weinstein's Blog The World Gets an Important New Drone That Can Save Lives — But Thanks to Our Politicians Not in the USA – Lauren Weinstein's Blog Separating the Chrome Browser From Google Could Be Terrible for Billions of Users – Lauren Weinstein's Blog In the War Between the Federal Government and States Against Drone Maker DJI, Americans Are at Risk – Lauren Weinstein's Blog The Website Age Verification Train Wreck – Lauren Weinstein's Blog How the U.S. Is Losing Its Lead in Technology and Science Research – Lauren Weinstein's Blog The Enormous Negative Impacts of the New Tariffs on the Technology Sector – Lauren Weinstein's Blog DOGE Is Destroying Social Security, and Seniors Are Already at Risk – Lauren Weinstein's Blog Time To Change Section 230? – Lauren Weinstein's Blog The Helpful Google Ombudsman (Who Doesn’t Exist) – Lauren Weinstein's Blog Commerce Department Proposes Yet Another Insane Chinese Drone Ban That Could Cost Lives – Lauren Weinstein's Blog AI Is Dooming Google, but Not in the Way Its CEO Believes – Lauren Weinstein's Blog [What say you, Spock?] My Proposed Terminology to Describe Bypassing Social Media Face ID Age Verification Systems – Lauren Weinstein's Blog Drone Hysteria and Bans Put Lives at Risk – Lauren Weinstein's Blog Australia’s Under-16 Social Media Ban Is Doomed – Lauren Weinstein's Blog DOJ’s Proposed Antitrust “Remedies” Against Google Would Be a Disaster – Lauren Weinstein's Blog Users have the most to lose – Lauren Weinstein's Blog “I Am the Very Model of a Google AI Overview” – Lauren Weinstein's Blog What Google Should Do About Their Search Generative AI Overview Answers – Lauren Weinstein's Blog Generative AI Is Being Rammed Down Our Throats – Lauren Weinstein's Blog Evil – Lauren Weinstein's Blog The Nightmare of Google Account Recovery Failures – Lauren Weinstein's Blog Google and Seniors – Lauren Weinstein's Blog About Google and Location Privacy – Lauren Weinstein's Blog
Why You Shouldn’t Use Google’s Chrome “Auto Browse” Agent...
2026-02-04 · via Lauren Weinstein's Blog

Various AI firms have launched so-called “AI browsers” and in particular what are called “agentic AI” browser features. And now Google has announced they’ve made massive AI upgrades to their Chrome browser which is by far the most used Web browser on this planet. And these Google Gemini AI features are becoming available to different classes of users paying or not paying over time, so you may not see some of them yet but you can feel pretty confident that eventually you will.

Frankly, I don’t recommend voluntarily using ANY current generative AI products from any firms. Google is indeed trying to push their Gemini AI into everything. But right now I want to warn in particular about what Google is calling Chrome “Auto Browse”. This is Google’s Gemini “agentic” AI system. And I’ll cut right to the chase at the start here, my very strong recommendation is, even more so than with other AI features, that you do not enable Auto Browse, do not use it, do not touch it. And I have the same advice for any other agentic AI systems from other firms.

What these systems do is in various ways take over your Web browsing. The AI literally masquerades as you, using your accounts and other credentials, and clicks its way around the Web to perform actions that normally you would do yourself. The concept is that in theory you could just tell the AI to find the best deal for something on the Web or book your vacation or clean up your duplicate photos or whatever, and the AI agent would run around and do all this for you.

I’m sure you already see why this has so many experts concerned, because we all know how AI systems spout misinformation and get confused, often can be manipulated in nefarious ways by hidden prompts on their inputs and so on. A three year old has more common sense than AI, because these AI systems have NO common sense. And we’ve already seen stories of people devastated when using these agentic AI systems when the AI deleted all their files or took other just awful actions.

Now, here’s the REALLY important part. It might be assumed that if these systems make terrible mistakes on your behalf using your accounts and credentials, that the AI firms would take responsibility. Well, think again. Google for example with their new Chrome Auto Browse pops a warning saying explicitly that actions taken by their AI on your behalf are YOUR responsibility. If the AI screws up, YOU get the shaft.

That’s the WHOLE ballgame as far as I’m concerned, and why I don’t recommend using agentic AI at all. These systems typically have settings that again in theory are supposed to let you control what sorts of actions they take, what files of yours they have access to and other parameters. Google’s for example at this point reportedly is supposed to stop just short of letting the AI click the final BUY NOW button creating a charge on your accounts. And of course they say you should monitor the AI’s actions.

This is all basically hogwash. Google must know that most people do not have the background or time to keep track of how these AIs are configured or what they’re actually doing, and if you have to monitor the AI to see if it’s messing up, much of the whole ostensible purpose is lost from the get go.

There’s a lot more technical detail of course. For example, your private browsing activities may be uploaded to Google as part of all this, triggering an array of additional privacy issues.

But as far as I’m concerned, this is a very straightforward decision. Even if Google for example were willing to accept responsibility for errors that Auto Browse makes that could potentially cause enormous problems for users — and AGAIN they’re refusing to accept that responsibility — I would not ever want these AI agents performing actions on my behalf — I won’t be using them.

If you’re willing to let these hallucinating Large Language AI models loose on your phone or desktop computer and let them go merrily clicking around the Web using your accounts and credentials, that’s your choice of course, but being a guinea pig for Big Tech AI isn’t anywhere on my personal bucket list.

–Lauren–