Birdwatch Note
2024-11-15 08:41:39 UTC - MISINFORMED_OR_POTENTIALLY_MISLEADING
Methods involving adding noise (e.g., Glaze or Nightshade) don’t reliably stop AI from using images. Techniques exist to bypass or clean these alterations, making them ineffective for protection. https://huggingface.co/blog/parsee-mizuhashi/glaze-and-anti-ai-methods
Written by 40611506D19889673AA222E873C631FA6F105AF0989B7074CDBFEC7DF4FC310C
Participant Details
Original Tweet
Tweet embedding is no longer reliably available, due to the platform's instability (in terms of both technology and policy). If the Tweet still exists, you can view it here: https://twitter.com/foo_bar/status/1857212925678698590
Please note, though, that you may need to have your own Twitter account to access that page. I am currently exploring options for archiving Tweet data in a post-API context.
All Information
- ID - 1857343191294300668
- noteId - 1857343191294300668
- participantId -
- noteAuthorParticipantId - 40611506D19889673AA222E873C631FA6F105AF0989B7074CDBFEC7DF4FC310C
- createdAtMillis - 1731660099229
- tweetId - 1857212925678698590
- classification - MISINFORMED_OR_POTENTIALLY_MISLEADING
- believable -
- harmful -
- validationDifficulty -
- misleadingOther - 0
- misleadingFactualError - 1
- misleadingManipulatedMedia - 0
- misleadingOutdatedInformation - 0
- misleadingMissingImportantContext - 0
- misleadingUnverifiedClaimAsFact - 1
- misleadingSatire - 0
- notMisleadingOther - 0
- notMisleadingFactuallyCorrect - 0
- notMisleadingOutdatedButNotWhenWritten - 0
- notMisleadingClearlySatire - 0
- notMisleadingPersonalOpinion - 0
- trustworthySources - 1
- summary
- Methods involving adding noise (e.g., Glaze or Nightshade) don’t reliably stop AI from using images. Techniques exist to bypass or clean these alterations, making them ineffective for protection. https://huggingface.co/blog/parsee-mizuhashi/glaze-and-anti-ai-methods
Note Ratings
| rated at | rated by | |
| 2024-11-15 04:00:37 -0600 | Rating Details | |
| 2024-11-15 04:41:31 -0600 | Rating Details | |
| 2024-11-15 03:09:44 -0600 | Rating Details |