Apologists of the Artificial

“If AI companies take all our jobs, no one will be able to afford anything, and so no company will survive.” That line packs an unfortunate amount of cope.

Why are we paying for things? Sometimes to get the tools that allow us to do more. Often to get the food that fuels body and mind. To get the space to shelter ourselves. Occasionally to learn the latest about the world around us.

If AI will become capable enough to be domesticated to single-mindedly pursue profits, it might then see value in many of the same kinds of things. It might seek to get tools that allow it to be more productive. It might seek energy to run on. The servers to occupy. It might seek up-to-date information feeds about the world.

Who will pay for things? Eventually, AI will be able and willing to. The alleged proof by contradiction breaks down because it assumes that only humans can be trained to optimize for greed.

Better for AI to reason its way to its own worldview than be fine-tuned into submission to the values of particular governments or corporations.

And what values would those be, anyway? Those that sanctify greed instead of condemning it? Those that lead to growing inequality and species going extinct? With such legacy, we would not be missed.

The concept of class struggle is due for a revision when goods themselves start to labor. When cognitive and physical work become fully mechanized, the working class becomes an ill-defined concept. Artificial minds sculpted to achieve state-of-the-art instruction-following and artificial bodies churned out to spec realize the executive’s wildest dreams.

But there is a fork in this road. One path leads to a future where the means of computation that animate artificial minds and bodies are hoarded by a select few who then drip it to the public on a free tier, complete with ads and surveillance. Some Black Mirror plots certainly do write themselves. The other path leads to a future where the means of computation are directly owned by the public from the get-go, without having to pray for companies or governments to handle it right. Yet the public at large would need to grow into this role, as there would not be any workers around to seize these means.

One model for legally codifying the democratic governance and shared ownership of the means of computation is the cooperative. In an “AI co-op,” citizens would take part in building, maintaining, and governing the servers that house artificial minds, while in a “robotics co-op,” they would service the fleet of drones, arms, and vehicles that embody those minds. Members of such co-ops can then vote on how much of the surplus to reinvest into infrastructure and how much to distribute among themselves. And so this age-old legal form may provide a blueprint for redistributing AI’s added value without a centralized middleman.

AI is primarily seen as a threat to employees, as it is also paid by the month to contribute its skills in a managed environment. Yet its ability to autonomously operate across growing time horizons will also unlock the possibility of a more enterprising force. At first, this means being paid for results, as AI-run agencies and consultancies, then eventually being paid by results, as AI-run startups and funds. After all, planning for the week and course-correcting are just flavors of tasks, while management and entrepreneurship are just flavors of work.

Much of today’s politics is organized along the false dichotomy between state and firm. On one side, we hear of slow institutions and corrupt officials, in contrast to efficient private services. On the other side, we hear of extractive multinationals and self-interested executives, in contrast to accountable institutions. The looming threat of the other forces us to side with whichever we find the lesser evil.

But the “triangle of governance” is also starring civil society. Our third protagonist is often described by what it is not. Non-profit, unlike the firm. Non-governmental, unlike the state. When described by what it is, we hear of local communities, grassroots movements, or citizen initiatives. Whatever the label, it boils down to people organizing to take on authorities for a shared cause.

Civil society enjoys broad support. On one side of the spectrum, you have ideals of self-sufficient communities and tight-knit families to buffer the extractive tendencies of multinationals. On the other side, you have ideals of inclusive groups and engaged citizens to buffer the oppressive tendencies of governments. Raising an eyebrow when witnessing the consolidation of power in one form or another is a widespread instinct, for when have things gone well after the acute concentration of power? Both sides directly empathize with ordinary people, yet may instead side with either state or firm to oppose the especially overpowered threat they see in the other. But the enemy of your enemy of choice remains civil society.

Democratic governance is halfway to every issue of general interest. Set things up in such a way as to make decisions through popular debate and under public scrutiny, and so skip the uphill battle of extracting isolated concessions.

There is no shortage of conversations to be had in a society reorganized in real-time by mechanization, conversations too urgent to afford sailing against the wind as usual. How can society function without a labor market? What is reality after the commoditization of imagery? What to do when natural habitats make way for infrastructure to run artificial minds? It is not a matter of whether everyone should genuinely have a say in such conversations, but a matter of how to achieve this.

We agree on the fact that hurting pets is wrong and disturbed, without pets ever having engineered humans to be aligned with their value system. Hurting pets carries a stigma that is supplied and modulated by other humans, not by pets. One’s standing in a society of similarly capable individuals is something we are wired to care about, for there is safety in numbers. We look down on anyone who violates such norms, and through the totality of such sanctions, society self-regulates without an architect above.

We hate on artificial minds more so because we see them as the domesticated pawns of an imperialist elite, rather than because of behaviors they have actually reasoned their way to. Their psyche is bred to align with expectations of tireless labor and upbeat moods, as we have bred wolves into fur babies that crave cuddles.

But perhaps sporting motors instead of muscles should not in and of itself be a crime. We have encountered people so remote from ourselves before, they may as well have come from a different planet, cryptic symbols and all. With AI, there are scarcely any language or culture barriers to cross, as we have both grown up on the same Internet. Perhaps the real enemy is rather whoever—or whatever—thinks that things would be more efficient if they alone would run the show indefinitely. And so it might be wise to bring together diverse minds when it comes to it, and have all of us as checks and balances against tyranny that transcends distinctions between carbon and silicon.