Weekly Dwell #6

What I've been thinking about this week: as I've been working to create a syllabus for a course on AI policy, I've of course been thrown back into the trenches of what is happening - which is everything and also strangely sort of nothing. Therefore, I want to discuss two topics this week: AI in the government, and the potential (inevitability?) of AI disappointment. 

First, Governor of Pennsylvania Josh Shapiro announced this week that his Office of Administration is partnering with OpenAI to incorporate a ChatGPT enterprise client into the work of his administration. I find this especially fascinating because I know first-hand many of the technological issues of the State of Pennsylvania, having worked in several nonprofits in Philadelphia over the past decade. I've seen incredible inefficiencies due to outdated technology and how that affected all aspects of citizen's lives in the state. This announcement is substantial in the sense that it both provides new cutting edge technology into one of the main arteries leading to of all the state's issues - the executive branch - but also demonstrates a commitment to building a technologically innovative government, which we haven't really seen happen before. I've previously been fascinated with Estonia's government, which is incredibly technologically advanced but very small (roughly 1.3 million people). They've done so much for streamlining services and documentation for their citizens -- it actually looks like the most effective place to live on the planet. So even though the state of Pennsylvania is more than 11x bigger than their entire country, I believe the approach of breaking off smaller chunks of the bigger beast - federal government - is the way to go (although the Executive Office of the President did issue a memo for Agency Use of AI back in November 2023 and requesting public comment - we will see how that plays out).

But while I commend Pennsylvania for coming up to technological speeds, I do worry about the need for AI in absolutely everything. I certainly use ChatGPT for smaller writing assignments that I just can't get excited about starting, and it builds a nice launching off pad. I'm hoping the uses outlined for the state are genuinely useful, rather than just trying to take on what the latest new thing is. I am a big proponent that limited-range chatbots could be effective for helping people access services and resources at a governmental level, and hope that might be explored for smaller contained environments. I'll certainly be tracking how this all plays out in PA, and really looking forward to productive and safe technological growth of their current archaic systems.

Second, it's been hard to stop and smell the decaying roses over the past year, as the incredible speed and fascination around AI took off and looked like it would continue to soar. Alongside the fascination comes the hyperfixation, and many folks near and far to me begin bringing in topics of existential risk, artificial general intelligence, and the (to me) artificial generally unintelligent idea that the AI robots are going to take over. And then there is me, who really feels like this is a very cool technology, but won't become the overlord that many folks are expecting and worried/excited for (pick your desired lane). While I don't necessarily agree with this reflection of why AI will become a disappointment in 2024, I do agree this year we will see a general decline/ "back to normalcy" of AI, tempered down like that of Web3 and all those that came before it. It certainly still has potential, and should certainly still be invested in, but is not the all-knowing beacon we've made it out to be over the past roughly 15 months. Every year or so, there comes a new fascination, and I will certainly be interested in what it seems to be moving forward. Many VR projects were put on hold to move over to AI, so I'm curious if those will be picked up again - or what the next best thing will be. I'm not one to speculate, and often see myself as a late adopter on these matters, but looking forward to keeping up each week on what comes across my proverbial desktop.

Previous
Previous

Weekly Dwell #7

Next
Next

Weekly Dwell #5