this post was submitted on 27 Aug 2025
85 points (96.7% liked)

Technology

74679 readers
2785 users here now

This is a most excellent place for technology news and articles.

Our Rules

  1. Follow the lemmy.world rules.
  2. Only tech related news or articles.
  3. Be excellent to each other!
  4. Mod approved content bots can post up to 10 articles per day.
  5. Threads asking for personal tech support may be deleted.
  6. Politics threads may be removed.
  7. No memes allowed as posts, OK to post as comments.
  8. Only approved bots from the list below, this includes using AI responses and summaries. To ask if your bot can be added please contact a mod.
  9. Check for duplicates before posting, duplicates may be removed
  10. Accounts 7 days and younger will have their posts automatically removed.

Approved Bots

founded 2 years ago
MODERATORS
L3s@lemmy.world
enu@lemmy.world
technopagan@lemmy.world
L4s@lemmy.world
L3s@hackingne.ws
L4s@hackingne.ws
85
Anthropic’s auto-clicking AI Chrome extension raises browser-hijacking concerns (arstechnica.com)
submitted 4 days ago by salarua@sopuli.xyz to c/technology@lemmy.world
8 comments fedilink hide all child comments
 

Security researchers have concerns that Anthropic's Claude for Chrome is vulnerable to malicious prompting. Claude for Chrome allows users to chat with Claude as they browse the web. Claude can read webpages, fill forms, and click on links and buttons to perform complex tasks for the user. But Anthropic's testing revealed that 11.2% of malicious prompting attempts succeeded even with safety measures in place. One test case was a malicious email that asked Claude to delete all emails in the user's inbox for "mailbox hygiene". AI researcher Simon Willison states that an 11.2% success rate is unacceptable for so-called AI agents, especially when several AI companies are releasing their own browser extensions. One competing product, Perplexity's Comet browser, was found to be vulnerable to a prompt injection attack that instructed it to start password recovery for the user's Gmail account. Although Perplexity attempted to fix the issue, Comet remains vulnerable to this attack.

Archive link

top 8 comments
sorted by: hot top controversial new old
[–] desmosthenes@lemmy.world 26 points 4 days ago

obviously lol

[–] Imgonnatrythis@sh.itjust.works 25 points 4 days ago (2 children)

First, I can't believe people are paying $100-200 / month for Ai crap. Second, if it were free or very cheap and I could sandbox it to only respond to painful cookie request menus to reject cookies, I would use it. I have consent o matic but it does a shit job and only works on a small percentage of sites.

[–] morto@piefed.social 24 points 4 days ago (1 children)

only respond to painful cookie request menus to reject cookies

You can do that just with ublock with the annoyance list, or using an extension like i don't care about cookies. Simple and efficient, no need for an "ai agent" for that

[–] Imgonnatrythis@sh.itjust.works 6 points 4 days ago

I don't care about cookies seems to be for people that feel that way. It sounds like if it has any difficulty blocking cookies is just accepts the to kill the pop up. I do care about cookies. It's my understanding unlock will just get rid of the pop-ups and allow cookies to default to accepted.

Both of these tools are detrimental imo as they just work against the protections that were attempted to be put in place by the EU to help minimize cookie abuse. Using these accomplishes Exactly what websites want. They have made pop ups and menus annoying enough that people are happy. To ignore them and allow them to default to aggressive tracking methods.

[–] Blisterexe@lemmy.zip 1 points 4 days ago (1 children)

Consentaumatic does the cookie thing

[–] Imgonnatrythis@sh.itjust.works 5 points 4 days ago (1 children)

On an extremely limited number of sites. It performs very poorly in real world use. Especially for academically oriented sites I've noticed.

[–] Dynamo@lemmy.world 1 points 3 days ago (1 children)

You can report websites in the extension if it does not work.

It’s still better than just accepting everything in my opinion. I trust that the coverage of websites/cookie prompts in the extension will improve over time.

[–] Imgonnatrythis@sh.itjust.works 1 points 3 days ago

I have been using for a long time. Yes it definitely is better than accepting everything! . It's a noble effort but it''s quite limited. An AI tool focused on doing this at least in theory could likely perform better.