I’m a groupie. My intelligence is somehow often dumbed down by the fact that if someone I respect and admire does/says something, I’m bound to follow them. Case in point: I bought Offscreen mag #17 because there was Jason Santa Maria in it (and also I know Offscreen mag and have read it a few times already, and I know I’m going to have a nice time reading it, of course — I’m not such a groupie, come on).
The good thing with reading a whole book of interviews is finding out about people you don’t know about, like Ashwini Asokan (co-founder of Indian AI startup Mad Street Den).
We humans have a tendency to polarise every discussion, exemplified in the “If you’re not with me, you’re against me” chorus. It’s kind of the same opposition when it comes to artificial intelligence. It’s either “They’re going to replace us”, Terminator-style, or somehow it’s magic that is going to fix everything. I loved reading what Ashwini has to say about that:
In reality, I believe we’re still far away from any AI that is truly meaningful and impactful, in part because our way of thinking about it is still very linear.
What we read about AI in mainstream media is mostly following the notion of superhuman intelligence. Think of headlines like ‘AI beats human at chess’. This line of thinking […] is very one-dimensional. There are many different models of intelligence, of cognition, of reasoning, and of problem solving because there are many different species. With most of these models we haven’t even started experimenting.
Right now, we are very focused on performance-based intelligence: the idea of doing the same things, but just better or more efficiently. Consider the path we’re on – the fact that we are literally destroying the planet that feeds us – it might be wrong to assume that as a species we possess the most advanced intelligence. What about social intelligence that allows some species to create more sustainable, self-preserving populations? What about emotional intelligence? […]
So, rather than thinking of superhuman intelligence, I’d like to shift the conversation towards systemic intelligence. Can we build artificial systems that are not independent entities but that connect various models of intelligence?
So much food for thought.