Description
One of the world’s leading AI companies, Palantir, is named for the scrying device in the Lord of the Rings trilogy that enabled the characters to see the future. This paper debates whether a Palantir can be created for or by Artificial Intelligence. It asks two core questions: 1.) can we use AI to 'see' the future, and 2.) can we anticipate possible AI futures, including how the technology might develop and the potential impact of its deployment. In answering these questions the paper examines how intellectual enquiry around these technologies can become more adept and rigorous at predicting and anticipating future outcomes relating to AI technologies, and, in particular, how IR can help drive methodological innovation in foresight, forecasting and futurology. This is particularly important when it comes to AI technologies because of the secrecy around their development (due to requirements of commercial protection or national security) and the opaqueness of development processes. Some of the most crucial and important questions in relation to AI, including whether an Artificial General Intelligence (AGI) or Artificial Super Intelligence (ASI) will emerge, are driven by uncertainty, which fuels fear and anxiety. Futures work has been done in other convergent disciplines – foresight and wargaming has become an influential tool for assessing security issues, for example. But this work has not been done for AI. Scholars have also bemoaned the waning predictive power of IR in general, including in relation to unforeseen events such as the end of the Cold War, 9/11, the Arab Spring, the Ukraine war and the current Gaza Israel conflict. There is thus an opportunity to refresh IR’s role and function in this area, which is arguably one of the disciplines defining features. The goal of the paper is therefore to explore how the discipline of IR can build a new futurology of AI, theoretically and methodologically, including by the application of futures methods such as foresight scenarios, back casting, wargames, simulations, sociotechnical imaginaries, AI-based forecasting tools and iterative psychological experiments.
