Kin.art is a new tool that aims to protect artists’ portfolios hosted on its site primarily by using image segmentation to prevent AI from poaching complete images. It is one of a number of tools that have recently entered the arms race between AI data scraping and content protection on the creator’s side, such as the University of Chicago’s Nightshade. At time of writing, texts remain vulnerable.
Optable, which offers data collaboration services for the advertising eco-system, has opened an Early Access Program for its activation capabilities based on its integration with Google’s Privacy Sandbox. This will allow advertisers to build and target audiences and publishers to activate and measure audiences, all without compromising privacy. Optable calls it an “easy button” for privacy-led advertising.
At the end of February, LinkedIn will stop updating lookalike audience data. It appears that the audiences will still be available for use but will essentially be static, thus quickly declining in value. It’s not clear why LinkedIn has taken this decision, but it is recommending Predictive audiences and Audience Expansion as alternative options.
Microsoft has confirmed that Reddit has blocked its Bing search engine from crawling content; Reddit seems to have blocked DuckDuckGo too. Reddit content will still appear in Google search results, but then Google has paid Reddit $60 million to allow it to train its AI models on that content. Reddit says the block has nothing to do with the Google deal: okay.