Systems Thinking: Untangling the Threads of AI Discourse
One of the things that the best developers I’ve ever worked with have in common is their reliance on systems thinking. We’ll likely be covering a lot of systems thinking topics as things roll forward around here, but a central aspect I want to talk about today involves the ways that system thinking encourages both breaking things down to understand the elements that make up a system, but then looking at how they come back together to form that system and make it different than just those contributing elements.
I’ve been thinking about a specific part of the “Commuting to the Future” presentation I gave at Minnebar on Saturday (still aiming to record a video version this week to share with everyone else and to get the version out that didn’t quite fit in a 40 minute session).
That part encourages people to look at discourse (including their own) about AI and pay attention to which things being said are about these different aspects of it:
The technology itself: What AI can and cannot do, the differences between versions and products
Structural applications: How organizations, governments, and economic systems use AI, their motives, incentives, and broader implications
Individual impacts: How the above aspects affect specific people we know and care about
When you look with those different lenses, you start to see how often people (including ourselves) are jumbling those things together, hopping between them as it suits our own agendas.
I’ve been thinking about that along with the HBR research I also included on the uses of AI that demonstrate a broad range of things, including many very personal uses, both for good and bad, that people are using it for. Because I’ve been seeing a lot of news items and anecdotal stories that push me to contemplate how all of this links together. I’m going to invite you to do the same.
With all of these things, they’re highly likely to spark immediate reactions, but also consider which lens triggers each response. Are there other lenses where the alternative perspective contradicts it or makes sense too? If pieces of the scenario were changed at one of those levels, would it be different? If your initial responses are all declarative statements about right and wrong, about how things ought to be, etc., what would it be like to look for questions you could ask that could make it look another way?
The first item was this Yann LeCun (Head of AI at Meta and NYU Professor) response about AI companionship. Mr. LeCun often says things that send me through the cycle I’m asking you to contemplate. I have an immediate reaction, but then I sit with it for a bit and see nuance. His statement that many of the objections to AI companionship also apply to pets is one of those.
I also saw someone I follow on Threads write a very vulnerable thread (hence not sharing it directly) about a mental health crisis that their AI tools helped prevent from turning into a spiral. There are lots of days where I see multiple of these kinds of stories, where people with one or another disability, mental illness, etc. and gets genuine help they were unable to get from their existing resources or the people around them. I also see people who respond, dead certain that those people are deluding themselves about it being helpful, on nothing more than a few hundred characters in the social media post of a stranger.
And, I also see articles about how people are using AI to pretty unambiguously go down a real spiral of confirmation bias and self-delusion in Rolling Stone’s “People Are Losing Loved Ones to AI-Fueled Spiritual Fantasies”.
There are, of course, no shortage of takes on all of this that come to definitive conclusions about it, what “needs to be outlawed completely”, what the singular unified purpose and motive of AI is, etc.
This web of things all have been stewing over the last couple of days for me. As they do, I look at them through those lenses. Outlawing all AI, a systemic/government solution, to completely end the overall tech itself, based on the uses of individuals of specific products made by specific companies is exactly the kind of thing I mean when I share those lenses and the fact that people often jumble, blend and hop between them.
I hope this gives you tools to examine these topics from multiple angles. I welcome thoughtful conversation but specifically ask that we avoid the reflexive, definitive judgments that characterize much of this discourse. Let's dig deeper and consider AI's implications from various perspectives.