Conversation

Notices

1. Electronic Frontier Foundation (eff@mastodon.social)'s status on Monday, 22-Jan-2024 08:31:43 JST Electronic Frontier Foundation

   Research into adversarial watermarking for AI is just beginning, and while there’s no strong reason to believe it will succeed, there are some good reasons to believe it will ultimately fail. https://www.eff.org/deeplinks/2024/01/ai-watermarking-wont-curb-disinformation

   • Roamin' Chemicals (roaminchemicals@mastodon.social)'s status on Monday, 22-Jan-2024 09:10:36 JST Roamin' Chemicals

     in reply to

     @eff Plenty of would-be scammers are technically unsophisticated and wouldn't take extra steps to remove or prevent watermarks. It's not a silver bullet, sure, but would it really be a failure if we could confidently filter out even 10% of machine-generated spam and disinformation with a quick check? It would be low sensitivity but high specificity; not ideal but still useful, like a rapid covid test.