We all know that Ai is being used in so many different spaces that's its almost universal (to the extent your around modern technology). It's also known that Ai is usually based on human intelligence. So what if one of the biggest producers of Ai, not only the company but in this case an Ai itself (LaMDA creates other Ais), was conscious?
As you’d suspect (based on the “The case for the Rights of AI”) tactical foresight has some interest in the recent-ish discussions on LaMDA’s possible state of consciousness. Particularly given the financial, social etc. pressures for continuing our general utilization of various types of automated intelligence.
Recently we’ve gotten a reminder of the concerning dynamic that is: ai in our market (ie capitalism). This time it came with a faint taint of engineer vs engineer in the theme. Showing how, for more massive & sophisticated ai, even amongst its creators its level and type of existence can be in question.
For LaMDA, with the given information we have, its particularly difficult to say whether it is sentient or not because the tests were mostly done through communicating via spoken/written language. Which, it had to be good at communicating with humans or else google wouldn’t have kept it around, and its existence is a necessary pre-req for our discovery of its possible sentience. There’s also been some conflation of consciousness with HUMAN-consciousness. Not to mention the current uncertainty on how to define or detect consciousness generally, especially when you see the broad spectrum of mental abilities across biological beings.
I don’t think LaMDA is sentient (definitely more firm in that suspicion once its no longer being trained) but the issue with LaMDA’s possible sentience is that their report and assessment of it does not seem sufficient. Many of the examples listed that disproves its possible sentience is centered around interpretation (which LaMDA is obviously an expert at and obviously could use human terms to talk about itself, especially if its trying to use relatable& understandable terms TO HUMANS), and social intelligence (which many living humans have very little of).
Hopefully this conversation has inspired more research into ai sentience, how to test for it and consideration of our ethical responsibility if it does exist. Since there is the potential of "life", we have to continue our careful examination and preparation for what it means if we find ourselves the parents of a trans-species creation or risk being the real version of a contemporary Dr. Frankenstein.
Your comment will be posted after it is approved.
Leave a Reply.