Artificial Intelligence (AO) is a broad and evolving field, with various definitions emphasising different aspects of intelligent software or machines/ robots. AI term was coined in 1955 by John McCarthy.

There is a good article at digitalwellbeing.org discussing popular definitions of AI from business and science. A lot of definitions evolve around AI being a set of computer science techniques that enable systems to perform tasks normally requiring human intelligence. Murat Durmus, CEO and Founder of AISOMA, talks about AI as the automation of activities associated with human thinking, such as decision-making, problem-solving and learning. He emphasizes the creation of machines that perform functions requiring intelligence when performed by people. Besides, AI is the study of how to make computers do things at which, at the moment, humans are better – read this article.

There is no universally accepted definition, but most researchers and AI developers emphasize AI’s capacity to exhibit intelligent behaviour and perform tasks traditionally associated with human intelligence.

Quite often, intelligence is defined as something that one can measure. There is a famous test measuring AIs.

The Turing Test, proposed by British mathematician and computer scientist Alan Turing in 1950, is a method for evaluating a machine’s ability to exhibit intelligent behaviour that is indistinguishable from a human.

Imagine a conversation between a human judge and two hidden participants – one human and one computer. The judge’s task is to figure out which participant is the human, and which is the computer by asking questions and analysing the responses. The test is conducted through text-based communication only. The judge can ask any questions or engage in any conversation topics. Both the hidden human and the computer try to convince the judge that they are human. If the judge cannot reliably tell the computer from the human, the computer is said to have passed the Turing Test.

The test has been influential in the fields of AI and philosophy, sparking debates about the nature of intelligence and consciousness.

While no machine has definitely passed the test as originally proposed by Alan Turing, there have been a few notable incidents where AI systems have come close or partially succeeded in fooling human judges.

In 2014, a chatbot named Eugene Goostman, designed to simulate a 13-year-old Ukrainian boy, convinced 33% of judges that it was human during a Turing test competition at the Royal Society in London.

ChatGPT 4 managed to fool human testers 40% of the time in two-minute conversations in an experiment when it was tested along with other large language models (LLMs). When prompted to make small spelling mistakes and quit if the tester became too aggressive, the AI succeeded in fooling testers.

As AI technology advances, particularly in natural language processing and LLMs, we may see more sophisticated attempts at passing the Turing test in the coming years. However, many researchers now question whether the Turing test is still the most relevant or effective measure of machine intelligence, given the limitations and potential for false positives or negatives in the test design.

A false positive occurs when a judge incorrectly identifies a computer system as human. A false negative happens when a judge incorrectly identifies a human participant as a computer. Both errors highlight a key limitation of the Turing test. It only evaluates behaviour, not true understanding or consciousness. Intelligent-seeming responses in machines occur without them comprehending the meaning.

We consider Alan Turing one of the most important scientists of all ages. He died by suicide in 1954, after years of persecution and mistreatment due to his homosexuality. He had to undergo chemical castration through a year of estrogen hormone treatment to avoid imprisonment for being homosexual. This ‘therapy’ had a severely detrimental impact on his mental and physical health, and he took his own life.

It wasn’t until 2009 that then-Prime Minister Gordon Brown issued a formal apology for Turing’s treatment.

Turing’s tragic fate highlights the severe discrimination and injustice faced by LGBTQ+ individuals in mid-20^th century Britain and serves as a sombre reminder of the human cost of such prejudice.

We invite you to subscribe to The Turing Lectures and The Turing Podcast delivered by The Alan Turing Institute in the UK. It is a treasury of information about AI and data science and an incredible sign of appreciation of our generations to Alan Turing’s scientific footprint – https://www.turing.ac.uk/.