LECTURE | Trystan Goetze (Cornell)

Title

Engineering Illusion: “Artificial Intelligence” and Conceptions of the Magical

Abstract

The escape artist and magician James Randi spent the latter half of his career debunking claims of real magical powers by purported psychics and faith healers, based on his observation that mystical charlatans and stage magicians each use exactly the same techniques of illusion. But for an illusion to truly mislead, the sensory experience must be accompanied by something cognitive: propositional attitudes constructed from the wrong concepts. In other words, for an illusion to be deceptive rather than merely fascinating, the illusionist must succeed in making the viewer think that the trick is, in some sense, real. A mendacious mystic is thus a sort of malicious conceptual engineer.

In a memorial to James Randi published in the journal Nature, science writer Philip Ball remarks that scientific advances don’t necessarily eliminate possibility for such credulity. On the contrary, some developments in science and technology can enable new forms of deception. One prominent illustration is recent developments in the field of so-called “artificial intelligence.” In this paper, I highlight several ways in which technologists encourage us to conceive of “AI,” each of which leverages representational devices which are continuous with forms of illusion and practices of magic and religion. Much of “AI” turns out to depend upon these forms of illusion for its appeal. I then reflect on the ethics of illusion, in order to distinguish cases where such illusions are harmless from cases where leveraging illusion shades into unethical manipulation and deception.

Zoom Access

Meeting ID: 654 5588 0974

Passcode: CEN24