Artificial Intelligence
Since apparently everyone is playing with DALL-E and Midjourney now and Google engineers are forming friendships with the machines they created, it looks like AI can really do more than just choosing the next song on our Spotify playlist.
Some thoughts on the narratives around AI:
1) “AI is becoming too human”
When an engineer claims that LaMDA, Google’s language model, has become sentient and newspapers ask whether “AI is getting too human for comfort”, we’re missing the point. It is not that AI is getting too human, we humans are becoming too machine-like. The “progress” of AI tells us a lot about our regress as humans, it shows us how we interact with each other these days: Indirectly, via electronic devices, with screens or speakers, mediated by algorithms. No human sitting in front of a blinking and noisy server rack would mistake it for another human. But looking at a screen and reading words that pop up, it is harder to determine who or what send them.
The question about what form and shape consciousness can take is an interesting one, but when the impression of consciousness is merely an artefact of the means of interaction, the question is a different one. We need to ask ourselves how much time we spend interacting with machines and how much time we spend interacting with humans, and how much time we interact with humans in a shared physical environment and how many interactions happen in a virtual environment.
This is relevant. How much of the online hate and trolling would happen, if people were talking to each other directly instead of anonymously typing out their frustrations at home? Less, as studies, experiments and common sense show. How much harder would it be to dehumanize and “other” someone else, if they were a breathing human being looking you into the eyes? How come we are connected to so many, yet it’s often loneliness that kills us? ( Some of these discussions revolve around McLuhan’s “the medium is the message” idea, check out this podcast episode to go deeper. )
Not just the questions change, but also the solutions. The solution is not another app, another platform or another algorithm to monitor the AI. It’s about building community, meeting people and talking with each other, moving and feeling our bodies.
2) “AI will destroy the world”
There are worries about AI going haywire: We give it a task and it uses any resource available to fulfil that task, regardless of the damage it causes along the way. That’s the old paper-clip machine example, or the story of ‘The Sorcerer’s Apprentice’ from Goethe and Mickey Mouse.
This is not a hypothetical problem, but it relates more to ideology than to AI. We have already invented and released a system that is maximizing towards a single objective and destroying life on this planet along the way: It’s called capitalism. We are currently sacrificing humanity’s future for the sake of private profits. Technology can help us accelerate the destruction, but it is not the cause.
3) “We have to win the AI race”
There are lots of fatalistic descriptions of an “AI race”: When China pours some billions into AI research, the US and Europe HAVE TO pour in even more billions to stay in the race, because losing that race will be really bad, like our economies imploding or something and China winning the world.
There is no race. We can stop at any time, reclaim our agency and make the game meaningless by not playing it any longer. We’ve been running in circles and have lost track of the goal and forgot the reason we started running. In the long run, success will be defined by building resilient communities, healthy ecosystems and societies that can adapt to a changing world, not by an AI trophy on the shelf that collects dust.
Here is a race I’d like to take part in instead: The Natural Intelligence race. Learning from billions of years of evolution to discover how life works and to regenerate this planet. Check out Leen Gorissens work at the Center for Natural Intelligence:
