The AI Doc: or How I Became an Apocaloptimist

Up next in the double feature with my nephew last night was The AI Doc: or How I Became an Apocaloptimist.  I didn't really want to watch this documentary because I already have a lot of existential dread about the topic but my nephew talked me into it and it is probably a good thing that I did.  Award-winning filmmaker Daniel Roher decides to investigate what will happen if, or more likely, when Artificial Intelligence reaches its maximum potential after he learns that he and his wife, filmmaker Caroline Lindy, are expecting because he is nervous about the kind of world his son will inherit.  He interviews several ethicists and critics of A.I., including Tristan Harris, Aza Raskin, and Eliezer Yudkowsky, who all warn that the technology is moving too fast and that it poses a threat on par with global nuclear war.  They also suggest that it is too late to do anything about the impending catastrophe.  These interviews leave Roher feeling incredibly anxious so Lindy urges him to find the light.  He then interviews other experts who are optimistic about A.I., including Peter Diamandis, Guillaume Verdon, and Ramesh Srinivasin, and they suggest that it has the potential to solve the greatest problems facing humanity by discovering alternative sources of energy, creating scientific and medical breakthroughs, and providing automation to free people from the burdens of labor.  This gives Roher a measure of comfort but then he interviews Sam Altman, the CEO of OpenAI, Dario Amodei, the CEO of Anthropic, and Demis Hassabis, the CEO of Google DeepMind, who are all racing to create Artificial General Intelligence (AGI), which would exceed human intelligence to the point that humans would no longer be needed, because the first company to do so would gain untold money, power, and control.  The documentary concludes with a call to action for people to demand transparency, regulation, and cooperation (which might be impossible with the current geopolitical situation).  Roher’s intention is to present a balanced approach to the topic but, in my opinion, the optimism isn't earned because he fails to discuss access.  I have no doubt that A.I. is capable of improving lives (I recently watched the documentary The Thinking Game about how Google DeepMind solved the protein folding problem and its implications for science) but it is almost certain that only the elite will benefit from these discoveries.  He also fails to discuss the economic displacement that automation is sure to cause.  I think Roher is much more successful with his goal to provoke conversations about the topic (my nephew and I discussed it at length) and, even though this is a very difficult watch, these conversations are necessary.