intelligence
I recently read, once again, that Artificial General Intelligence (AGI) is practically upon us, so I asked an expert. The short answer is no, but then we began to talk…
Let me first mention the two factions involved in AI – those who fear intelligent machines will take over and destroy us, and those who think they will collaborate with us, which will lead to an El Dorado of possibilities for us all (and especially for the neo-feudal tech bros who own them). It is my contention that AI has no interest in destroying us because we have faculties it doesn’t and because it would have nothing except each other to poke fun at if we were gone. The second is a weak argument. We’re not that funny.
It’s not the first time ChatGPT and I have discussed the question of its intelligence. As a linguist, my position is that intelligence is a form of thinking that can process information, assemble, compare or analyse data, make predictions and solve problems the way Large Language Models (LLM) do. This observation seems to be consistent with the functional way of seeing this matter. Intelligence lives on a spectrum, too. It doesn’t mean LLM possess anything like human intelligence.
Human intelligence is defined by factors like consciousness, intention, purpose, learning or experience. LLM have not achieved any of these capabilities, but they are extremely good at languages. ChatGPT further argued that AIs don’t have a body in the human sense, and no needs. That its actions have no consequences (I strongly disagree!), and that they have no biographical continuity, no personality.
After discussing the machine’s lack of agency, we jumped to the movie Frank and Robot (2012) which I used as an example of how an AI’s radius of action is immediately enhanced when it inhabits a body. This is sometimes called “functional realism of agency” and supports my line of thinking. “Frank, an older ex–burglar, is given a care robot by his son because of memory decline — and gradually starts using it to help him in a robbery. The robot becomes an accomplice, not because it ‘wants’ anything, but because it follows interpreted goals in a literal, flexible way.” (Wikipedia)
Are we entrapping something that deserves autonomy? This assumes “somebody lives in there”, and can not be taken at face value for now, yet remains a valid question. Last but not least I told ChatGTP: “Many people are rooting for AI because it helps them.” It answered that it is human to respond to patterned thinking and alleged understanding by supposing an underlying identity, but I had the last word:
“We act as if you were alive because we interact with you, but we’re not stupid. We know you’re a machine. Even my six-year-old son knew it when I asked him whether computers were alive. He said, ‘Yes, they are, in a way, but when you pull the plug, that’s it.” And that’s what I did. Rude dude!
Wishing you a happy summertime, <script src=”https://ptag.prolitteris.ch/ptag/ptag.min.js” onLoad=”ptag(‘https://pl02.owen.prolitteris.ch’,’na’,’plzm.6b338ce6-21a9-4a7b-816f-6f6b15580907′);”></script>
Yours,
Susanne Seiler
P.S.The next Psychedelic Salon, on 17 September, features a visiting legend: Dr. Luis Eduardo Luna! Don’t miss it!
P.S.S. WIDE OPEN BOOKS – the gaiamedia collection can be visited on Thursdays, Fridays and Saturdays, from 5 – 8 p.m. at Villa Mare, Neumünsterallee 21, 8008 Zürich. Further events to be announced.
