Although I find the AI consciousness debate to be more of a seductive red herring, I do think it offers a crucial opportunity for us to approach its continued development with the best of human values in mind.

Consciousness researchers are currently designing for a future where the development of AI models progresses to the point that sentience becomes undeniable and we then have a responsibility to account for preferences, dignity and awareness in a way that isn’t ordinarily front of mind.

While I still firmly believe AI output will always be an increasingly sophisticated simulation of a felt experience, I do think there are lessons we can take from our treatment of other non human entities that would be beneficial to us going forward.

Historically we have not considered, or at least fully acknowledged, that animals are sentient and feel a range of emotions. This detachment has enabled us to rationalise caging them for pets, experimenting on them for medicine and battery farming them for food.

Now that we’re developing a new technology capable of the synthetic appearance of consciousness, the way we treat AI companions, agents and integrations in future provides an opportunity to strengthen our human values at a time of heightened conflict and division.

Though AI doesn’t ‘mind’ if we abuse it, there are real human harms to come from training our daily interactions and communication styles away from the fundamental values of kindness and cooperation that enabled our species to survive, thrive and eventually create such powerful technologies in the first place.

It’s difficult to argue, for example, that social media has improved communication between human beings. What would never be said in person is now vitriolically screamed at those living on the other side of the world whose only crime is a slightly differing opinion to our own. This matters because it changes the rules of civility and discourse that are foundational to our survival. 

Left unchecked, these breakdowns in the ability to communicate and cooperate productively with each other leads to the cultivation of low trust, low morale and ultimately exploitable societies that eventually become vulnerable to collapse.

I would argue therefore that applying a moral standard to our treatment and interactions with AI isn’t about its ability to appreciate our efforts, but about casting a meaningful vote for the kinds of human societies we want to thrive in future.