In-Person
Seminar

False Portents: The Limits of Artificial Intelligence for Strategic Bargaining among States and the Mirage of Free Knowledge

Harvard Faculty, Fellows, Staff, and Students

The advent of sophisticated Artificial Intelligence (AI) systems has provoked vociferous debate among pundits and scholars over their likely impacts on strategic bargaining under anarchy. Techno-optimists and techno-skeptics disagree over the breadth of qualitative transformations that AI is likely to inject into world politics. Still, both concede that at least some aspects of tactics, strategy, and grand strategy will be irreversibly transformed by the ability of AI to leverage big data and big compute to extend the frontiers of human knowledge and human control. In this seminar, the speaker will argue that this expectation is largely unwarranted.

For more information, contact susan_lynch@hks.harvard.edu

The Oracle of Delphi Entranced
The Oracle of Delphi Entranced

Speaker: Anatoly Levshin, Postdoctoral Fellow, International Security Program/Technology & Geopolitics 

The advent of sophisticated Artificial Intelligence (AI) systems has provoked vociferous debate among pundits and scholars over their likely impacts on strategic bargaining under anarchy. Techno-optimists and techno-skeptics disagree over the breadth of qualitative transformations that AI is likely to inject into world politics. Still, both concede that at least some aspects of tactics, strategy, and grand strategy will be irreversibly transformed by the ability of AI to leverage big data and big compute to extend the frontiers of human knowledge and human control. In this seminar, the speaker will argue that this expectation is largely unwarranted.

The ability of any intelligence to apprehend, navigate, and control its environment depends upon its capacity for articulating and revising defeasible abductive judgments about the causal structure of that environment. There is, at present, no known way to automate this process, which leaves AI agents entirely dependent upon their human operators for the supply of adequate abductive judgments. Whenever these judgments succeed in exposing the underlying causal structure—as, for example, in physics, biology, or board games—AI can productively augment human capabilities. But the domains of tactics, strategy, and grand strategy are fundamentally different in this critical regard: here, our stock of abductive judgments is cruder and sparser. An environment that remains causally opaque to its human inhabitants is also, and for precisely the same epistemic reasons, opaque to AI. There is no shortcut to free knowledge. 

In the fullness of time, the portents of AI transforming strategic bargaining under anarchy may well come to pass. But that will only occur in the wake of qualitative growth in the precision and range of our abductive judgments about tactics, strategy, and grand strategy. For the foreseeable future, then, those portents remain false.

Admittance is on a first come–first served basis.  Tea and Coffee Provided.

 

Up Next