Meta will signal when political ads are created by AI

Meta, the company responsible for the social networks Instagram and Facebook, announced this Wednesday (8) a new measure to combat misinformation on the platforms due to content created through artificial intelligence (AI).

Meta’s new policy for advertisers Those responsible for paid posts must notify users that AI-altered posts have been “digitally created or modified” from any vehicle.

The measure applies to: Audio, video or photos showing real people “doing something they’ve never done”, posts imagining a real event happening differently, or content showing people or events created from scratch by AI.

Target against election disinformation

Company’s goal To prevent the spread of fake content, especially during the United States presidential campaign.

Deepfakes by candidates, for example, of speech or actions that lead them to say something controversial or get into embarrassing situations, are already circulating on social networks to damage or praise politicians. In these cases, there is almost always no warning that the content was created by artificial intelligence.

According to the target, Subtle and aesthetic digital changes do not need to be reportedsuch as cropping, color adjustments, or changes to the sharpness of the content. Besides, notification will be added to Meta ad libraryThis means that new attempts to boost the same content or the same post will have the settings already saved in the system.

Measurement should only be released in 2024 and may not apply to all countries at the same time, but Meta is expected to provide more details about the new rules later.

Recently, the company itself launched a generative AI: the Llama 2 language model, which is expected to become a chatbot for mobile phones next year. It has also created a platform that can translate text and audio into up to 100 languages.