Meta has decided not to offer its new AI advertising tools to political advertisers. These tools, which include features like image adjustment, background generation, and automatic caption writing for videos, are part of Meta’s expansion into using machine learning for advertising. The decision comes ahead of an anticipated contentious national election, and Meta has not updated its advertising standards to reflect this choice publicly.
Other social media platforms have similar restrictions, with TikTok and Snap banning political ads entirely, and Google using a keyword blacklist to avoid political content in its AI tools.
However, Meta does provide exceptions to its rule, allowing AI-generated video content that is misleading, as long as it is for parody or satire. This policy is currently being reviewed by Meta’s independent Oversight Board, following a case where Meta did not remove an altered video of President Biden, arguing that it wasn’t AI-generated.
In July, Facebook and other tech giants agreed to voluntary commitments proposed by the White House to implement safeguards in AI development. These include enhancing efforts to detect and eliminate harmful AI behaviors, sharing safety information within the industry and with the government, and creating a digital watermarking system to verify authentic content and distinguish it from AI-generated material.
Key Takeaways:
- Meta has introduced a set of generative AI tools for advertisers that can create backgrounds, adjust images, and write video captions.
- Meta will not allow political marketers to use its generative AI advertising tools to prevent misuse during election cycles.
- Meta and other tech companies have committed to implementing safeguards in AI development, including adversarial machine learning and digital watermarking.
“The experimental features can adjust images, generate backgrounds and automatically write video captions.”
More details: here
