Researchers say it took fewer than 20 prompts for a public AI tool to find a flaw (now fixed) allowing anyone on a Zoom call to hijack another participants’ device.

Leave a Reply

Your email address will not be published. Required fields are marked *