Modern Use of Alan Turing’s Imitation Game

(L-R) Keira Knightley, Matthew Beard, Matthew Goode, Benedict Cumberbatch, and Allen Leech star in THE IMITATION GAME.

The Turing test was believed to be the one plus ultra-state of artificial intelligence. The AI pioneer, Alan Turing, used it for the first time in 1950 to develop a benchmark for progress. It measures how quickly human thinking ability is reached or simulated by AI. However, it all comes down to the question of whether an AI program can sufficiently deceive people to make you think that this machine is a thinking person and not a machine. According to http://botpress.io/blog/turing-test-2/, Turing also spoke of an imitation game, adapted by modern AI development systems. Here, the candidate communicates in natural language via a computer with a human and a machine.

Undoubtedly, AI systems have become increasingly “smarter,” embedded with sensors and bodies that can move autonomously in robots and vehicles worldwide. AI programs can now write articles about sporting events, talk to customers, or perform operations. Dan Rockmore, director of the Neukom Institute at Dartmouth College, now wishes to transfer the Turing test to the field of art with a new competition, as announced. The competition is based on the Turing Test in Creativity by definition of art, which it creates with imagination and skill, expressing important ideas or feelings. Something should be submitted from AI programs that can write a short story, a poem in sonnet form or a DJ mix from which people are convinced that they are not produced by a machine but by a human being.

The question of whether something can be understood as art, about the good and incomplete work, is not avoided. After all, it is not just a question of quality, that determines whether the work’s appearance confirms if it was created by a person, or a “soulless” and “uncaring” machine. The distinction is certainly not less a problem, but it is assumed that what a talented person creates is supposedly imbued with inspiration. Here, one might ask whether it is an individual’s humanity that is expressed in a work of art.

The Neukom Institute inspects production conditions. For the poem (PoetiX) and the short story (DigiLit), the program takes 24 hours to create something in the way a New York editor or a Master of Fine Arts collection theoretically can make it appear. This is significantly less time than an average author needs to write something so high in quality. The judges will receive a collection of works in which computer-generated projects come from distinguished authors but know that they should not be able to prevent how they are recognized by their style.

The short program should run on a normal computer with less than 8 gigabytes of RAM, a word that governs immediately, and a virtually unlimited number of original short stories can write with high probability. Submissions must provide the source code. For the DJ mix (Algorhythm), the AIprogram has two hours to create a 15-minute piece based on a list of 1,000 songs. Human DJs do the same under the same conditions. The pieces are played for voluntary dancers, who are then asked which mix probably comes from machines.

Applications for this competition need to be submitted by a specific date. Then a reward of $5,000 will be available for each computer-generated submission held by the judges as a human-made work of art. Judges at Rockmore don’t believe this will be possible in the case of the short story or the poem. But there will also be prizes of $3,000 for all submissions, showing the improvements, like a computer handles the idea of creation, even if no one is deceived by it. Finally, the piece being interesting is not enough for the piece to be considered a success, it must be exceptional in technique and application.

Be the first to comment

Leave a Reply

Your email address will not be published.


*


− eight = two