Pedro Domingos

It's alarming that NeurIPS papers are being rejected based on "ethics reviews". How do we guard against ideological biases in such reviews? Since when are scientific conferences in the business of policing the perceived ethics of technical papers?

Hi Pedro, I helped create the NeurIPS ethical review process. Looks like there's a healthy discussion going on here already, but let me know if I can answer any specific questions. Up front, I should say that the ethical reviewers gave feedback; they did not accept/reject papers.

Hi Raia, my point is that papers are being rejected based on this "ethical review", and I object to that. The papers' ethics should be evaluated by the readers, not pre-emptively by what is effectively a censorship board.

How would readers address a lack of informed participant consent, ethically questionable methodologies, or ends? By the time of publication, the damage is already done. “An experiment is ethical or not at its inception; it does not become ethical post hoc” - Beecher 1966

As your quote says, that is to be decided at inception time, not publication time.

Then let’s have an IRB for machine learning that uses data about humans. You must agree with that?