🤖 AI browsers are NOT safe!
There is a thing called "prompt injection" and it works.¹
Funnily the thing that most see as a major issue with AI, the
crawling of the #web and one-way use of it's content, is exactly what makes their AI browsers unsafe.
If you place malicious #code in that very content, the AI scans it & then runs it² on your OS 🤯
This issue has been known to the #tech corps for years³, but they released their #AI browsers nonetheless 🤑
1/2
RE: https://mastodon.social/@madeindex/115446644659078400
🤖 #OpenAI just officially admitted that they will never be able to make their #AI Browsers truly safe!
Of course they won't let themselves be stopped from selling their #product by such a minor detail ;)
"We expect adversaries to keep adapting. Prompt injection, much like scams and social engineering on the web, is unlikely to ever be fully “solved”..."
https://openai.com/index/hardening-atlas-against-prompt-injection/
#chatgpt #atlas #it #news #internet #browser #artificialintelligence #ki #promptinjection #tech #technology #software