Researchers from Queen’s University Belfast are worried AI “ghostbots” could start rising from the dead without UK privacy laws to stop them. These digital specters can happen when AI is used to create digital reincarnations (chatbots, deepfakes, or holographs) of people who have died. Problem is this might terrify actual live people (assuming robots haven’t gotten rid of us before this happens).
The law requires companies to post signage near store entrances to notify customers when biometric data is about to be collected.
We haven’t written much about Adobe’s entry in the generative AI footrace because their main focus has been Firefly, a tool for creative departments. But they made a splash last week with an announcement that they’ll indemnify Firefly users against claims of copyright infringement, which is a big concern. The news was probably over-hyped; Adobe’s actual announcement is carefully circumscribed: “enterprises also have the opportunity to obtain an IP indemnity from Adobe for content generated by certain Firefly-powered workflows.”
AI-edited content is less concerning than AI-generated content — that’s the takeaway from Meta’s adjustment to its AI labeling on Facebook and other platforms. The “AI info” label will continue to appear on AI-generated content. Where content has only been modified or edited using AI, the label will appear more demurely in the post’s menu.