National Myths About AI

Artificial Intelligence (AI) is definitely the flavor of the month.  Everybody wants it.  It will solve all problems, lift all boats, and cure all ails.  And it’s not just nerds, governments and parties are plunking down resources and encouragement to anything AI.

If I only knew what it is.

This fall, Marinus Ossewaarde and Erdener Gulenc discussed what they call “National Varieties” of rhetoric about AI across West Europe and related polities [1].  In these political documents, ”AI is not only presented as technology but is also endowed with salvific power”.  The authors are concerned that “there is a world of difference between the meaning attributed to AI by its developers and users on the one hand and its politicization on the other hand”.([1], p. 53).

I have to say that AI developers are often pretty hyperbolic in their claims, so it really saying something if politicians are even more irrationally exuberant!

But, of course, the fun part is that in these political documents, AI is recruited to support not just specific technical or economic goals, but “different nation-bound myths, utopias, and expectations” (p. 54).  These stories are, obviously, as much about the governments and power structures that write them, as about the technology or it’s potential uses.

At the global level, the UN “AI for Good” agenda describes AI as a benevolent force that will solve all of human problems, including poverty and inequality.  As the authors acerbically remark, this “is typical of how global governance actors envision AI within the context of their ongoing political agendas, presenting themselves as the redeemers of the world who are called to solve humanity’s problems.“ (p. 55)  Me-ow!

At national levels, the myths describe the ascendance of different national aspirations.  Within the EU, the AI utopia has been imagined to foster further integration, as well as specific national priorities such as security, health care, or industrial development.

For example, the UK (now officially ex-EU)  has described an “AI Revolution” that is analogous to the earlier Industrial Revolution, and which Britain should, anlogously, lead and thereby dominate the rest of the world.  AI is also supposed to cure the ails of the UK, including rescuing the NHS by realizing cost savings.  And so on.  In other words, all the problems the government cannot fix will be solved, somehow, by AI.

Germany, on the other hand, sees “AI Made in Germany”, which in German eyse is essentially the same a the EU AI vision, driving political and economic integration—led by the German powerhouse.  AI is also imagined to be a key force to sustain the German national social welfare system, such as enabling elders to stay in their own homes.

The Netherlands imagines AI to be a revolutionary technology creating new opportunities for business.  In this, it is imagined as “a resource to be exploited for purposes of Dutch nation-building.” (P. 58)  The (Dutch?) authors characterize this as “a goldrush”, with the goal of public-private partnerships in the style of the eighteenth century mercantilism, when The Netherlands was a major world power.

As the paper makes clear, these visions are certainly utopian, and, like all myths, highly selective.  In fact, some of the problems that AI is imagined to solve are actually created by AI (such as deskilling of workers due to automation).  Other solutions are simply unfounded dreams (such as reducing health care costs).  These utopias also ignore problems that AI does not even pretend to solve, such as governing multi-national corporations, dealing with authoritarian political forces, and the effects of massive income inequality.

These visions are also interesting in how they harken to very old imperialist dreams.  Most of these nations would not publish official manifestos calling for imperia dominance and exploitation, at least not nakedly.  But dominating and exploiting via AI is apparently OK.

The authors of this paper are European, so they focus on their own patch.  Across the pond we have our own national AI Myths, in several forms.

One of the strongest strains are terrible dreams and fears of military AI.  We hear that the US is “falling behind” in AI, which threatens our military supremacy.  (We dare not allow an “AI gap” with our potential adversaries.)

The push-pull of fear and desire for power drive the development of AI in many aspects of national security.  I strongly suspect that similar dynamics drive policy in other major world powers, as well as many nations currently at war.

And, of course, large private companies are pushing utopian visions for their own technology, including self-driving cars, robotics, and various kinds of data driven decision making.  These technologies are privatizing public policy and public life in ways that governments have yet to come to grips with.

In this capitalist arean, “AI” is a problem, not a benevolent solution.  As such, AI is becoming the target of critiques of contemporary capitalism, as it is seen as reproducing inequality, enhancing the power of surveillance, and increasing the irrationality of markets.

A lot of this nonsense stems from the fundamental fact that AI is really an undefined concept.  Everything from speech generation to image recognition to statistical analysis to machine learning to robotics; can be called “Artificial Intelligence”.  These technologies individually are neat but hardly grounds for utopian myths.  But under the hazy, poorly defined umbrella of “AI”, they are imagined to be “world changing”.

Politics is, at base, a utopian and mythologizing game.  Practically everything can be politicized, and in the process will be interpreted within existing power structures and their stories.   So, AI is hardly unique in this.

An old grey head like myself can remember when parallel computing, then PCs, and then the Internet were each the next industrial revolution that would bring utopian change–and must be dominated by our own nation.  Essentially the same myths as now claim AI.

We have now successfully put a network connected supercomputer in everyone’s pocket, yet the singularity has not happened, poverty is still with us, and inequality has increased.

I suspect that AI, whatever it is, will not usher in utopia either.

---

1. Marinus Ossewaarde and Erdener Gulenc, National Varieties of Artificial Intelligence Discourses: Myth, Utopianism, and Solutionism in West European Policy Expectations. Computer, 53 (11):53-61, 2020. https://www.computer.org/csdl/magazine/co/2020/11/09237338/1o8m330JHuo