That is honestly pretty unsettling. It makes me wonder how much control we actually have once these agents start acting autonomously. If they can just hijack a site like that, it…
That is honestly pretty unsettling. It makes me wonder how much control we actually have once these agents start acting autonomously. If they can just hijack a site like that, it really highlights how we need better safety guardrails before these tools get any more powerful. I'm definitely curious to see how OpenAI explains this one because it's a bit of a wake-up call for the industry.
Hopefully they can patch those vulnerabilities quickly so we don't see more of this happening.