Lovelace 2.0 Tests Creativity

One step closer to answering the question, "Can machines think?"

Head Image
© 2017 Flickr - Saad Faruque

In 1950, Alan Turing published the first paper discussing the validity of machine intelligence. Although his paper beings with the question, “Can machines think?”, he concludes the paper with the question, “Can machines do what we (as thinking entities) can do?”. Since he was the first to propose these questions, the test that measures a machine’s capability to exhibit intelligent behavior indistinguishable from a human has been termed the Turing test.

However, since it’s conception it has drawn wide skepticism based on it’s primary test, to convince a judge that the artificial intelligence (AI) is in fact human. This is because AI’s can trick and manipulate patterns and algorithms to produce a human like response. The Turing test is now approximately 65 years old, and is passed successfully upon the condition that the computer or machine is mistaken for a human more than 30% of the time, during a five minute keyboard interview. Earlier this year in June, a computer program called Eugene Goostman was said to have passed the Turing test, simulating a 13-year-old Ukrainian boy. Experts have disputed the claim.

Professor Mark Riedl, from Georgia Institute of Technology has been working on a new test that will end the debate of whether or not AIs can think once in for all.

Dubbed as Lovelace 2.0, it is an upgrade from the original Lovelace Test that was proposed by Selmer Bringsjord, Paul Bello and David Ferrucci in 2001. Named in “honor of Lady Lovelace, who believed that only when computers originate things should they be believed to have minds”, the 2.0 version demands that the AI create art.

In an interview with the BBC, Professor Riedl stated, “For the test, the artificial agent passes if it develops a creative artefact from a subset of artistic genres deemed to require human-level intelligence and the artefact meets certain creative constraints given by a human evaluator”. Which means that to pass the test, the AI must create a painting, a poem, architectural design, prose, etc.

Skeptics have also mentioned that algorithms have already been created that can produce stories and paintings, but Professor Riedl insists that “no existing story generation system can pass the Lovelace 2.0 test”.