There is no AI in PANIC

 a friend recently sent me  https://www.nytimes.com/2023/02/26/opinion/microsoft-bing-sydney-artificial-intelligence.html in response to this output from a WIP at http://ai.numina.org

I responded with this:

No problem. I mean, I have been working in the digital for over 40 years. Already during one of my business trips to Salt Lake City in the 80s, an enforced and extended separation from my family, I felt used by evolution as a handmaid to help bring this new species to life. The AI built into video games at the time was extremely simple and illusory but people still bought it because we have evolved a theory of mind to interpret, predict and control the behavior of our own species and we project it onto anything which seems to be making decisions by a process we can't inspect directly.

The article had some worthwhile points to make, though when the author said, maybe we aren't talking about the right things, I assume he means opinion writers in mainstream publications and their readers because the potential risks of this technology have definitely been covered in the technology press and are widely discussed by the groups developing the technology though of course there are differences in how the seriousness of risks is accessed and how to address them.

I think the biggest error the author made is in his assumption that the obvious shortcomings, the transparent errors will just be fixed and then AI will continue its inevitable march to world domination. In fact many of the smartest people on the planet have been working on these issues at Google and they kept their work in the lab because they knew about the issues. OpenAI decided to go public with theirs even though they knew there were issues and Microsoft decided to make a product for the general public out of it, so Google demoed their version. In doing so the giants of the tech industry lost hundreds of billions of dollars in market valuation in days because of the inadequacies of he technology.

I am interested in using it for art because I think the very issues that make it not useful as a business tool make it interesting as a way to think about humanity, the obscurity and fallen combined with unexpected beauty. It is after all only replicating satistically what it has been trained on which is the combined knowledge of humanity as represented by texts on the internet. 

My advice for how to think about this technology is, without panic. In terms of dangerous technology I believe AI more resembles writing than H-bombs. Sure, gather capitalism and all the usual suspects, but democracy has the potential of aligning the interests of technology and society, so it is important to become informed and be thinking seriously about the issues and how they might be addressed politically. Rather than condemning a tool as witchcraft we should locate evil in the only place it exists, the hearts of men, and focus our attention on how to keep this tool from being used to nefarious ends.