Meta Expands AI Image Labels to Curb Election Misinfo
Meta will expand efforts to identify and label AI-generated images on its platforms ahead of upcoming elections globally, seeking to curb misinformation and deception.
Summary
Meta will label AI-generated images from major AI companies like Google, OpenAI, Microsoft etc. in addition to content created by its own AI tools.
These labels will appear in all languages available on Facebook, Instagram and Threads in the coming months.
Meta will work with other AI companies on technical standards to detect AI-generated content even without embedded signals like watermarks.
Users can voluntarily disclose AI-generated audio/video since those are harder to detect automatically. Penalties may apply for undisclosed deepfakes.
Prominent labels will be added for AI content posing high risk of public deception on important matters.
This effort to curb election-related misinformation builds on previous actions after issues in 2016 and 2020 US elections.