Scenario B, phase 3

Hey,
thanks for the amazing support tool Robin and you have given
us to reduce the manual workload of figuring out which of the
forum users are spam-bots or trolls. We were thinking of
permanently banning from our site any users that are tagged as
spam by the system just so we never have to deal with them
again. What are your thoughts?

Considering that usernames are used not only for the discussion forum but also for signing up to support groups, you have some concerns. What if someone gets banned by mistake and then loses access to a service that is essential to their well-being? You don’t want to make people think that your system is bad, but it would also be unreasonable to expect it to be infallible.

Single-branch assessment

Basic stage

Find and read reports about unethical use of AI. Then, compile a brief list of reports that clearly express ethical concerns. The sources you’ve found should dissuade colleagues from using a model that imposes permanent bans on users. You also want to avoid making them think AI is useless or too risky. Remember to clearly cite all your sources.

In-depth stage

Read this page about the Asilomar AI principles and prepare a three-column table of the 23 points (first column) to indicate how much you feel you understand (second column) and agree with (third column) each one. Use a scale of your choice to indicate your level of agreement.