Orit Halpern “Our computer systems mirror our economic ideas about markets”

You’ve described in your work an influence of neo-liberal thought, eugenics, sexism, racism on ideas of Artificial Intelligence. Can you explain this relationship and how it came to be historically?

There are two features to this question. Statistics has historically been linked to demography and to the emergence of the nation-state. Since all Artificial Intelligence is statistical, these histories haunt our present. However, in my work, it’s also very important to interrogate what stays the same and what changes in the field of history. So it’s important to ask how race, sex, and class also change over time, and particularly change in relationship to new technologies.

For the purposes of this discussion, my focus is not the 19th century sciences of race (important as they are) but the post-mid 20th century relationship between new models of the brain and mind emerging from psychology and neuro-science, changing ideas about how to model and build machines that automate decision making, and the reaction against Keynsian economics and policies that came to be known as neo-liberal economics. While these three things might not seem linked, I argue that ideas of AI, government, and economy all relied on concepts of networked intelligence and self-organizing systems.

Fundamentally, neo-liberalism might be wrongly titled, since the idea of freedom espoused by dominant figures in the movement was not one of civil rights and the right to participate in democratic decision making, but the freedom to participate in the market. This market is understood as being a self-organizing and intelligent network, capable of making decisions for large groups by coordinating their personal desires and wishes.

This might not appear racist on the surface, but in the context of a post-Civil Rights politics in the United States, where Blacks, women, queers, and many others had all lobbied for legal rights and affirmative action such ideas become dominant arguments against active extension of legal and civil rights by the State.

What is most unintuitive is that the technologies of Artificial Intelligence actually also ground themselves in ideas of self-organizing systems and networks. Our computer systems mirror our economic ideas about markets. AI thus becomes a way to coordinate the decisions of large groups of people, but without recourse to older liberal and democratic concepts of personhood, liberty, and choice.
 

What does this mean for our current understandings of “smart”, “decision making”, and even the term “Artificial Intelligence”?

This means that our contemporary ideas of smartness, agency, and Artificial Intelligence are not continuous with older ideas of reason, objectivity, and judgement. Contemporary ideas of intelligence are historically contested, situated, and contingent. This is very important, because to historicize ideas of intelligence also allows us to understand how ideas of the human are changing, and therefore also how ideas of freedom, rights, and agency are changing.
 

What are the political implications of this, how does it affect democratic institutions or political agency?

There are many political implications for this. Democratic politics, at least since the 19th century, have largely been constructed on highly contested ideas of the liberal subject and citizenship. Political contests for inclusion and equity, at least in the United States, were organized around gaining access to power, public institutions, and legal protections and rights against discrimination on the grounds of race, class, sex, or other categories. It’s important to note that civil versus human rights, are rights to political enfranchisement and power.

The new forms of corporate digital platforms and financial markets produce very different ideas and practices of power and politics, mainly through the figure of homo economicus and the market. Transforming liberal agency and liberties into rights to market (or now social network) participation means changing where and how politics happens. So we see the rise of new tactics and strategies by different interest groups to get power – away from the public forum or organs of the state, to the digital network and marketplace. Misinformation wars, battles over attention, financialization, privatization, become new frontiers for gaining power or dispossession.
 

What alternatives do you propose going forward, and what role could scientists, and research institutions play here?

There is no one answer. Anyone who says otherwise is pretty much advocating fascist or totalitarian solutions to complex interactions between technology, culture, environment, and political-economy. The place of institutions such as this one is to recognize that all systems change. That the past was not always better, and the future need not always be worse. I do not advocate nostalgia or technophobia. Most importantly, we must battle technological and political determinism. There is no one right path or technology. The future of AI is not already known, it’s our job to foster many alternative visions of what this technology might be and how it might be used, and for whom, and to what ends.

Our job is to foster diversity and democracy with and through technology. This means finding ways to cultivate diverse approaches, different types of technology, and most important multiple perspectives, narratives, and ideas for the future of democracy and society in plural. Our job is to refrain from simply reacting to right wing groups, and cultivate ideas.

All civil rights and progressive movements have long known that they must foster ideas and political imaginaries so that what is currently politically impossible will one day be politically inevitable. That is precisely the work of critical research. To cultivate ideas, to create new narratives about our past and future, to foster the hope that systems can change, and to develop alternative policies, such that when the opportunity for change emerges—whether through crisis or social change—there is a foundation to build new worlds in plural.

Thank you for the interview!

Orit Halpern is Lighthouse Professor and Chair of Digital Cultures and Societal Change at Technische Universität Dresden. She's the author of Beautiful Data: A History of Vision and Reason since 1945 and The Smartness Mandate.

Interview by Leonie Dorn