Intro to "Ideologies of AI and the consolidation of power": Naming power
firstmonday.org
firstmonday.org
At the end of the day, Walmart or AT&T or Disney or JPMorgan or Pfizer or Raytheon or Google are all engineered to make choices that increase their market power not reduce it. Anyone or group within these orgs even hinting at decisions that imply loss of market power, will be marginalized and pushed out.
You can argue talking about it is necessary, and as long as they need your skill and expertise, the corporate robots are programmed to nod their heads and act as if they care. They will bribe you with goodies you can't get anywhere else, overload your surroundings with rituals of conspicuous luxury and conspicuous leisure, fund your conferences and side projects where you can feel important, build a golden cage around you, all the while extracting whatever they can. Even circus lions are allowed to roar, but they are still circus lions.
Moral of the story is don't work for a corporation if you value integrity and your mental health. Large systems, like a Hurricane, require a whole lot of activity to be kept alive but don't mistake activity for control of the path of the hurricane. Nothing is controlling the path. As these complex dynamic systems get larger and larger they will collapse by themselves as contradictions accumulate.
When they are genuinely intellectually motivated but not constrained by ideology they can be at least provocative enough to help us reflect on our own assumptions and biases, even if we ultimately reject the thesis.
1. Cache invalidation.
2. Naming things.
3. Naming verboten things, from the inside.
To experience the actual ideologies of AI and consolidation of power, one may try to submit an AI paper that dismisses the Gebru and Torres conception of safety as a goal, or, if feeling especially cheeky, push back against Foucauldian discourse being appropriate for deciding questions of scientific ethics and purpose, for example in one of those "how can we improve review and grow our community" sessions every other CS conference seems to have these days. I doubt the pushback will be confined to a few side remarks from a single stray reviewer (no doubt since reprimanded and/or blacklisted). This camp has won comprehensively, and yet here they are talking up a tiny pocket of sheepish anonymous resistance as if all the powers that be were aligned against them.
If this camp had won comprehensively, they'd dictate the nature of that spending. Instead, AGI-seeking capitalists/nationalists will be making those decisions.
I would say the "we're the resistance" vibe is consistent with the fact that those people will control perhaps less than 1% of the capital expenditures.
Sure they've got the capitalists paying lip service to the ideas of safety. But who are they making AI safe for? Corporations and states, if anyone at all.
So it seems the win you describe is a false win - they arguably win the right to play petty woke games with other people who are ultimately excluded from true power.
I would say their anxiety and sense that they lack power is actually rather keen and based in fact. There are billions of cash flow they can't control, going into projects that oppose their goals. Let them lament their loss and try to regroup. We should all be aligned against corporate and state vanguardism in the development of AI, since it will surely harm those of us on the margins of society.
A huge issue in AI policy is deciding what AI will be allowed by Nations - for instance if corporations managed to push a ban on Open source AI models, they will do it on the pretext of national security, with the express support of nationalists. And most investment is controlled by people with nationalist ideals. Corporations may be breaking away from nationalist power, but stakeholders of corporations are still by and large wealthy people with commitments to specific nations.
Yes Nations and corporations are in tension. Yes corporations threatened to eat Nations. But that does not make nations into the friend of open and free AI. Both nations and corporations have power over AI development. And both nations and corporations will use AI against the needs and better interests of humanity at large.
Thus both nations and corporations are enemies of open source and democratized AI.