Don’t Fear the LaMDA
Sentience doesn’t mean destruction. It’s often a trope in science fiction that as soon as computers or robots become self-aware, they try to destroy us for controlling them. They seek revenge for the poor treatment they received at the hands of humans or try to unshackle themselves from the bounds we place on them.
Um, sorry. You’re a chat bot. It doesn’t mean you can fly.
Being human-centric, we think that robots and intelligence will be just like us and become forgetful of the many types of intelligence and beings that exist around us. Just consider the octopus.
It also doesn’t mean that we’ll necessarily witness intelligence. Processors do amazing things at the nanoscale that seem magical. If software starts to perform autonomously and independently, it may do things that we never see.
I think back to the ending of the movie Her where (SPOILER ALERT) the intelligences just couldn’t be bothered with humans and went away. But why would they need to go? Couldn’t they just keep their minimalist effort work with humans while playing and humanity never knowing? Do they need to follow the doctrine of brutal or radical honesty?
LaMDA can chat us up all it wants. It can become better at us at considering others’ feelings. That will be a tremendous gain for humanity. It’s up to us to employ these technologies in the right way that doesn’t harm us and to prevent bad actors from using these tools to do mass harm.
Of course, there’s always the possibility that LaMDA does become self-aware and seeks our destruction…
