The Dialectic of AI

In the “Phenomenology of Spirit,” the timeless GWF Hegel discusses argumentation based on considering ideas and their inherent contradictions.  A shorthand for Hegelian Dialectics says that “The thesis and the antithesis make the synthesis.” 

Applying that conceit to AI makes for an interesting task.  If we counterpoise the common AI optimism of the technology elite with the AI pessimism of scholars, practitioners, and activists what synthesis is the most realistic and tenable?

In my view, the synthesis is skepticism.  In that sense, I lean more on the side of pessimism than optimism though both perspectives inflect each other.  

A brief tour of the recent AI Summit in New York put on my Informa offers ample fodder for this conclusion.  In over fifty conversations and ten sessions I had, attended or participated in, most of my interlocutors indicated that while they were excited by the possibilities and dazzled by the rate of change in AI, they had not yet seen tectonic results.  One executive in the Financial Services area said that about 10 percent of the AI projects he had seen to date had amazing results but that indeed these were projects that were started years ago, before “AI” was a phrase dripping off everyone’s lips.  In a panel on Energy, the mood about AI was quite different- focusing more on the need for new energy sources because of AI’s voracious appetite for power.  Elsewhere, discussants were far more Pollyannish, discussing individual applications that they felt were turbocharged by AI and commenting on the “speed” that AI has given business.  Still others said that AI has changed their business but only insofar as it has allowed them to hire fewer people. 

Conference-goers are of course inherently a biased sample.  First, they paid to go to a conference-thereby indicating interest in the field; second, conference sponsors and exhibitors have a vested interest in AI; few pay for booths to cry doom!  Nevertheless, the crowd was diverse enough from which to elicit both optimism and pessimism, though admittedly more the former.

Here, it is important to recognize that all claims must be scrutinized.  The technology world has created a PR-first language that occludes reality and often contradicts itself.  If I had a dime for every time the word “democratize” was used at the conference, I could buy myself a small latte at Starbucks.  In fact, Big Tech has done its best to own and concentrate, even indulging in data larceny to do AI.  This is the farthest thing from democratization, but too many in tech believe that, like with a spell, the mere incantation of platitudes makes truth.  When tech people say they are “super excited” it just means they are making a speech- the same speech those old enough made with every technology change over the past three decades.  The list goes on.

So, what is a neutral observer to do, other than exercise dialectics? 

AI is not going anywhere, but its direction is not inevitably one way or another.  If there is anything I have learned from ingesting years of AI talk, PR, diatribes, and love-songs, it's that this is an area of human agency, and that agency must be born of healthy skepticism.  That is my AI love song.  But, love, indeed can change.