Mosseri urges AI-free labeling as C2PA stalls
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Adam Mosseri (Instagram head) said in December that fingerprinting real media would be "more practical than fingerprinting fake media" as AI improves.
- C2PA (content credentials standard) has been adopted by Meta and supported by Adobe, Microsoft, and Google, but its implementation has been "wholly ineffectual" despite broad industry backing.
- Not by AI offers a suite of badges for creators, requiring at least 90% human contribution, yet its verification relies on visual inspection and AI detection without independent auditing.
- Proof I Did It uses blockchain to issue immutable "Made by Human" tokens, and Thomas Beyer says this creates a market for verified human-made art with mathematically guaranteed authenticity.
- UC Berkeley experts (Jonathan Stray, Nina Beguš) highlight that defining "human‑made" is difficult because hybrid content blurs the line, and current detection tools are unreliable.
- Proudly Human admits it cannot prevent fraudulent use of its label but offers consumer verification tools and threatens legal action against bad actors.
Why it matters: Human creators stand to gain market trust and protection against AI-generated competition if a reliable, verifiable label emerges, while platforms and deceptive actors lose credibility and revenue when fraudulently applied badges erode consumer confidence; the push also pressures regulators to define 'human-made' standards and could shift industry investment toward blockchain-based verification, reshaping digital content authentication.
