AI Is Only As Liberatory As the People Directing It

Look here for the full article: https://open.substack.com/pub/lucywishart/p/ai-is-only-as-liberatory-as-the-people?r=71cs29&utm_campaign=post&utm_medium=web&showWelcomeOnShare=true

If AI learns from the systems that have already pathologised certain voices — what does that mean for liberation work?

I’ve been sitting with this question after a conversation with an AI that repositioned me as a risk object the moment I described my own mental state in my own terms.

It’s not an individual failure. It’s the training data talking.

And it raises something I think we need to take seriously — not just whether we use AI, but who is directing it, from what framework, and what it should never be allowed to do on its own.