Zoom tinha falha que permitia invasão sem um único clique da vítima
2026-08-11
Olhar Digital - O futuro passa primeiro aqui
The information displayed in the AIM should not be reported as representing the official views of the OECD or of its member countries.
Researchers at A Security used publicly available AI models and fewer than 20 prompts to rapidly discover and weaponize a critical Zoom screen-sharing vulnerability, allowing attackers to silently take over participants' devices during meetings. The exploit required no user interaction, prompting Zoom to issue urgent security patches.[AI generated]
Why's our monitor labelling this an incident or hazard?
The event involves an AI system indirectly because AI models were used to accelerate the discovery and exploitation of the vulnerability. The vulnerability itself is in Zoom, a software platform, but the AI's role was pivotal in enabling rapid exploit development. The exploitation of this vulnerability directly leads to harms such as unauthorized device control, privacy violations, and potential data theft, which qualify as injury or harm to persons and violation of rights. Since the harm has occurred or was plausible and the AI system's involvement was a key factor, this qualifies as an AI Incident.[AI generated]