Articles on the History of AI
- The brief history of artificial intelligence: the world has changed fast — what might be next?
- A (Very) Brief History of Artificial Intelligence (PDF)
"The Imitation Game" aka "The Turing Test"
The "Imitation Game" is a straw-person argument authored by computing pioneer Alan Turing, who posited that exploring AI by defining "intelligence" would be semantically difficult (if not impossible), and he instead presented a counterpoint thought experiment published in the Journal Mind in 1950, "Computing Machinery and Intelligence" (Mind, issue 59, pgs 433-460, 1950). He made the case that the mimicry of a human, when communicating only with text responses, is sufficient to practically define an "intelligent" machine.
The Imitation Game sets up an interview scenario in which an interviewer has tightly constrained access to two interviewees, through exchanging only slips of paper with them while sitting in isolated rooms. He suggests an truly "intelligent" computing system could fool the human interviewer (who would merely have these short text exchanges) into believing it itself is human. Turing explores a number of subtleties, arguments, and cases around this scenario made in his essay.
- Read the Wikipedia article here.
- The DOI link to the original paper here.
- Direct link to UMBC's PDF OCR of the original paper (has some scanning errors).
- Direct link to McGill Universities PDF scan of the original paper.
(Note that this essay has statements you may see as controversial or offensive. Rather than defend this as "just the way things were," I'd recommend reading the paper as a historical artifact rather than an endorsement of such a position.)
The "Turing Test" is arguably not as productive a thought experiment today as it was when originally posed, but it does raise questions still relevant. It's somewhat curious that the AI tools since 2022 getting the most attention are delivered as chat interfaces.
Turing Machines
Alan Turing made significant and foundational contributions to our understanding of computing systems and of computation, perhaps the most notable of which is the so-called "Turing Machine," which is a reductive model of what constitutes a fully capable computing machine.
If one can prove equivalency of a given system to the simple axioms of a Turing Machine, one has proven such a system can compute anything that any other complete computing system can. We say such a system is "Turing complete." And some counterintuitive languages, like the latest versions of CSS (cascading style sheets) are indeed Turing complete.
- A fun visualization of a Turing machine at https://turingmachine.io/.
- A formal discussion of the Turing Machine can be found here at Stanford's Encyclopedia of Philosophy.
- An interesting lecture from Stanford CS on the origin of the Turing Machine.
