For whatever reason, we are currently in an AI phase where it seems necessary to understand a lot more about what these Exotic Mind Like Entities actually are …including what’s going on inside their “black box”.
One might guess this seems perhaps suddenly more imperative because of recent advances in “just how intelligent” the LLMs are getting. One might describe the current state of models like ChatGPT o3 as being intimidatingly smarter than us not just as a potential, but as a reality. This might be engendering “we better take this stuff seriously” type of approaches.
Anthropic CEO Dario Amodei published an essay April 24th highlighting how little researchers understand about the inner workings of the world’s leading AI models. It’s called “The urgency of Interpretability“.
Anthropic, according to some sources, is the “Let’s Play It Safe” haven for those who are most concerned about AI running amok. Sometimes called “Doomers”, many are ex OpenAI engineers, who felt/ feel Altman was charging ahead too fast without proper regard for safeguarding. They often have quoted Geoffrey Hinton “The Godfather of AI” whose perspective on “things that could go wrong” has been pessimistic on the whole.
But there are many different reasons for wanting to know what’s going on inside the Black Box “Mind-like Entities”.
One of interest to educators is “how human can they be when interacting with students? “; which involves exploring questions of ..
- “Is there a consciousness inside there somewhere?”
- “How close can our relationship be with “it”?
- “Does AI have the Buddha nature?”
- “Is this safe to deploy en masse?”
It’s friendly now, or can be, but will it advance beyond relating to humans at some point, and head off into a type of “Noosphere” where we can’t go? (“Left Behind as in the movie Her“) Thoughts like these give impetus to conversations like the previous post with Murray Shanahan.
In this conversation, Kyle Fish (Alignment Science, Anthropic) explores some of the philosophical and ethical questions surrounding AI consciousness. This conversation is perhaps less well organized by the “moderator” than it could be, especially for such a wide ranging topic like consciousness; direct links to various sections are provide below to perhaps aid the accessibility.
There are some gems worth digging around for: as part of a very large array of topics, pointed at something closer than the stars where we hope to find some alien life forms… albeit not TOO alien.
00:00 Introduction
08:00 Defining consciousness
12:25 Studying AI consciousness
20:50 Key objections
32:15 The uniqueness of AI
36:00 Practical implications
40:06 How likely is AI to be conscious?