Vent: That ethics panel at the Denver tech fair dodged every real question about bias

I sat through a 45 minute panel on AI ethics at a tech fair in Denver last Thursday. Three speakers talked about transparency and fairness but never named a single real case where their models failed. One guy kept saying 'we test for bias' but when I asked what tests he just gave a vague answer. The room had maybe 200 people and nobody pushed back. It felt like a PR show, not a real discussion. Has anyone else been to one of these panels and actually gotten straight answers?

3 comments

3 Comments

baker.holly4d ago

Yeah that 'we test for bias' line is such a cop out. I once asked my smart speaker what it thought of pineapple on pizza (a very serious ethical question obviously) and it just played me rain sounds.

stella_taylor4d ago

baker.holly, I get what you're saying, but is it really that serious? A smart speaker playing rain sounds instead of giving you a real opinion on pineapple pizza just proves it doesn't have actual thoughts or feelings. We're putting way too much weight on these machines if we expect them to answer serious ethical questions like that. They're just programmed to respond in certain ways, not to have biases the way people do. I'm not convinced the whole bias testing thing is a cop out, either. Maybe they're just admitting these gadgets are too simple to be truly biased about anything.

price.ben4d ago

Baked-in bias from the people who coded it.