Third parties can use user content to train AI systems, even though Bluesky may not be doing so like other social networks
A machine learning librarian at the AI company Hugging Face extracted one million public postings from Bluesky using its Firehose API for machine learning research, according to a report by 404 Media.
They then uploaded the dataset to a public repository. Because of the ensuing controversy, Daniel van Strien later deleted the data, but it’s a helpful reminder that whatever you submit publicly to Bluesky is, well, public.
Although it is up to other parties whether or not they respect those decisions, Bluesky stated that it is investigating methods to allow users to express their consent preferences externally.
“Bluesky will not be able to enforce this consent outside of our systems,” the company wrote on social media. Outside developers will be responsible for adhering to these settings. We’re now speaking with engineers and attorneys, and we anticipate sharing more information soon!
Bluesky’s rapid rise to prominence will undoubtedly subject it to the same scrutiny as other major social media platforms.
