Meta's ban extends to all content, including non-paid posts, but exceptions are made for parody or satire. This policy is currently under review by the company's independent Oversight Board. In July, Meta, along with other leading AI companies, agreed to voluntary commitments set by the White House to implement technical and policy safeguards in the development of future generative AI systems. These include efforts to root out bad model behavior, sharing trust and safety information, and developing a digital watermarking scheme to authenticate official content.
Key takeaways:
- Facebook's parent company, Meta, has decided not to make its generative AI tools available to political marketers, especially in the context of the upcoming national election cycle.
- Meta's decision aligns with other social media platforms like TikTok and Snap, which ban political ads, and Google, which uses a "keyword blacklist" to prevent its AI tools from engaging in political speech.
- Meta does allow exceptions to this rule, permitting "misleading AI-generated video in all content, including organic non-paid posts, with an exception for parody or satire." This policy is currently under review by the company's independent Oversight Board.
- Facebook and other leading AI companies in Silicon Valley agreed in July to voluntary commitments set by the White House. These commitments include technical and policy safeguards for future generative AI systems, such as expanding adversarial machine learning efforts, sharing trust and safety information, and developing a digital watermarking scheme to authenticate official content.