Unethical Chatbots

The Global Forum on the ethics of AI took place recently in Kranj. Reading through the literature I was interested in the UNESCO recommendation on core values, particularly one on transparency:

“The ethical deployment of AI systems depends on their transparency & explainability (T&E).”

This relates to my “3 Laws” article last year.

https://paulhcleverley.com/2023/10/26/the-three-laws-of-data-management/

I would argue that AI systems like OpenAI’s ChatGPT and other domain focused Large Language Model driven chatbots are not ethical if they do not allow the user to be able to click on answers and see exactly the provenance of the information used to generate the answer.

This may be even more important in domain verticals where users of such systems may be using systems for research and decisions that ultimately have an impact on science and people’s well being.

These guardrails for transparency build trust in use of AI systems.

https://www.unesco.org/en/artificial-intelligence/recommendation-ethics

The Oxford Institute for Ethics in AI https://www.schwarzmancentre.ox.ac.uk/ethicsinai and Oxford Internet Institute https://www.oii.ox.ac.uk have some great papers in this area.