Forschungsseminar: Lecture by Dr. Ian Jennings, FHNW School of Life Sciences

The problem of AI alignment is usually framed as the question of how to ensure that increasingly capable systems behave in accordance with human values.

This talk asks a prior question: what if the greatest obstacle to alignment is not the machine, but our inability to agree on the nature, source and authority of the values with which we hope to align it?

Note

To ensure that the research seminars are an enjoyable experience for everyone, please bear the following in mind:

• If you arrive late, enter the room quietly, close the door gently and take your seat.
• It is fine to eat a light snack discreetly, but please do so without making any noise or distracting other participants.
• Please give the guest speaker your full attention and do not use your laptop or phone.

Thank you very much.