1. Can you stop companies from training generative AI using your data? No, not currently. 2. Is this dataset meant for training generative AI? 🤷‍♀️ but more likely for research and statistical analysis. 3. Is it ok to duplicate and distribute people’s data without agency to opt out? I’d argue no. - ThreadSky

mariaa.bsky.social • 115 days ago

1. Can you stop companies from training generative AI using your data? No, not currently.
2. Is this dataset meant for training generative AI? 🤷‍♀️ but more likely for research and statistical analysis.
3. Is it ok to duplicate and distribute people’s data without agency to opt out? I’d argue no.

Reposted from Daniel van Strien

First dataset for the new @huggingface.bsky.social @bsky.app community organisation: one-million-bluesky-posts 🦋

📊 1M public posts from Bluesky's firehose API
🔍 Includes text, metadata, and language predictions
🔬 Perfect to experiment with using ML for Bluesky 🤗

huggingface.co/datasets/blu...

Comments

mariaa.bsky.social•115 days ago

3. Do researchers duplicate and distribute data all the time? Yes and maybe we can finally talk about this and come to a consensus. It’s comes up a lot in ethics reviewing but the current community norm is to distribute copies. I’d prefer links or IDs a la the old Twitter API, despite downsides.

mariaa.bsky.social•115 days ago

4. Likely outcome from piling on and sending abuse to people making open tools? Companies aren’t going to stop taking your data, researchers internal to companies aren’t going to stop analyzing your data, but it will all be done quietly and without users’ knowledge.

mariaa.bsky.social•115 days ago

5. What is “AI”? This is a big umbrella term used to refer to generative AI like ChatGPT, statistical analysis for public health and social sciences, machine learning used to create small tools like “trending topics,” and much more. These carry different ethical tensions, not good to lump together.

khakhalin.bsky.social•115 days ago

At this point we need another word to replace AI, and it's painful. In most cases one can say LLM or ML and be fine. But there are rare cases when ppl are really working on _actual_ AI in the sense of designing intelligent systems. And for them, there's no "clean name" now! So weird.

khakhalin.bsky.social•115 days ago

This dataset in particular is clearly primarily ML for social science. But like, even these these terms got splash-contaminated by association. "Ah, you want to do science on me??? Inhumane experiments? Chips?" Strange times, strange times...

mariaa.bsky.social•115 days ago

And people may very good reasons to object to being experimented on! But the conversation is muddled and will be more effective all around if we can be precise.

shan23chen.bsky.social•115 days ago

I’m terrified by the massive openreview data. Potentially gonna bite back on us 🥲😥

mariaa.bsky.social•115 days ago

Is there specific stuff in that data you're worried about?

shan23chen.bsky.social•115 days ago

More on future potential reliance on LLM agent doing reviews and audits

Comments

Posting Rules

Reply