back
5 comments
I hate it when people criticize the Turing test as imprecise or unproductive. While it may be not be an important benchmark in serious AI research, it's been incredibly thought provoking and inspiring for over half a century. Many great scientists give us value beyond their publications in journals, essays, and dreams that help to put the bigger picture into perspective.

Also I'm glad the movie came out so it's easier to explain to my family about one of my personal heros.

Turing is also one of my heroes too, but that doesn't mean I have to worship every idea of his. As you pointed out the problem with the Turing Test is that IS as imprecise or unproductive.

This would be fine if decades of serious AI research and millions of dollars hadn't be wasted trying to fool humans with these stupid psychological tricks. I know first hand as I took an "AI" course at my University centered around their home grown chatbot. And it's not just my school, here's a list of 400 research papers dedicated to the topic.[1]

As the article points out Turing's idea that was never fully flushed out. It's nothing more than a distraction for researchers and a silly parlor trick for everyone else. Stong AI is never going to come by tricking the interviewer that you're a 13-year-old Ukrainian boy.[2] It's going to come by a fundamental understanding of the nature of intelligence.

[1] https://www.chatbots.org/papers/

[2] http://motherboard.vice.com/read/how-a-computer-beat-the-tur...

I don't think it is surprising at all that Turing took the possibility of telepathy seriously. Many scientists and educated people took the possibility seriously in the 1950s, 60s, and 70s. This is reflected in the many research projects conducted at the time in academia and government (most infamously, the MKULTRA CIA program).

You can see this belief represented in the "hard" science fiction published at the time, much of which treated telepathy and other ESP subjects seriously--including Bester's The Demolished Man, Asimov's Foundation series, Heinlein's Stranger in a Strange Land, Clarke's Childhood's End, and Niven's Known Space series.

There was a real and broadly shared sense, for several decades, that psychology and ESP might be the area of the next fundamental scientific breakthrough.

It was not until the 1980s that it became clear ESP was a dead-end, and today it's mostly mocked by the scientific community as a source of scams. But to apply today's mindset to Turing's 1950 paper is a pretty substantial hindsight error.

Telepathy exists, in the form of brain waves that can be measured and stimulated via electrodes.
Very interesting read. If anyone is interested in chatbots or conversational AI, I'm working on the space (in the open) superscriptjs.com