Bots are like hunters: there are good ones, those that provide information from a reliable source (for the weather, a good quip, a plane’s schedule, etc.), and bad ones, those that spread disinformation or act in the service of a rogue state to destabilize other countries. In which category will we place the army of bots that Meta is preparing to unleash on Facebook and Instagram? Hundreds of thousands of bots have already been created by (human) users of social networks, via an AI character workshop.
On Facebook, no one knows you’re a bot
This tool was launched this summer in the United States, and Meta plans to expand its availability in the future. However, most of these bots are not public. "We expect these AIs to gradually integrate into our platforms, like user accounts," says Connor Hayes, vice president of product for generative AI at Meta.
These bots "will have bios and profile pictures, and they will be able to generate and share AI-powered content on the platforms," he hopes. This is a "priority" for Meta, which wants to make its applications "more entertaining and engaging" over the next two years, including more social interactions with AI.
Content creators can already create AI avatars that answer their subscribers' questions while respecting the personality of the influencer. So tomorrow, will Facebook and Instagram only have bots interacting with each other in a vacuum? This is probably not Meta's intention, and it will also have to find solutions to avoid the risks of massive disinformation that these bots could generate.
The group indicates that AI-generated content should be clearly labeled as such by users of these tools. As it stands, it is very complicated to sort the true from the false and automatic tagging systems do not give the most convincing results.
Source: FT
0 Comments