There is a widespread assumption that the continued deployment and incremental optimization of AI tools will ultimately benefit everyone in society. But perhaps it is time to critically interrogate the value and purpose of using AI in a given domain in the first place? In 2021, political philosopher Annette Zimmermann published an essay for Boston Review called “Stop Building Bad AI”. In this conversation with Boston Review editor Matt Lord, Zimmermann expands on many themes from this essay, offering a clear overview of the key social and political questions that philosophers are addressing in the face of AI-related problems, such as algorithmic injustice, lack of democratic accountability for powerful corporate agents, and the kind of learned helplessness that results from coming to see AI development as inevitable.

ANNETTE ZIMMERMANN is Assistant Professor of Philosophy at the University of Wisconsin-Madison and a Technology and Human Rights Fellow at Harvard University. Their research focuses on the ethics and politics of artificial intelligence, machine learning and big data.

MATT LORD is senior editor at Boston Review. He studied literature, mathematics, and philosophy at MIT and Harvard and is editor, most recently, of Thinking in a Pandemic: The Crisis of Science and Policy in the Age of COVID-19 (2020).

Matt Lord (ML): One way of telling the story of philosophy's long engagement with artificial intelligence is to see a transformation from an earlier focus on logical, epistemic and metaphysical questions – “What is AI? Is it even possible?” – to a growing discussion of ethical, political and social questions. To put it crudely, there was a time when the field was once little more than footnotes to John Searle's Chinese room argument. Now, following significant advances in machine learning and computing power over the last decade, we have this well-defined new subfield of philosophy, the “ethics of AI”, to which a lot of scholars, including you, are contributing. What explains the change? And what does it mean to think about AI from within the perspective of ethics and political philosophy, in particular?

Annette Zimmermann (AZ): One of the reasons for this is that there is a growing body of work, which is identifying problems of clear normative – not just technical – significance, coming from computer science and applied statistics, as well as applied mathematics.