Artificial Intelligence and Machine Learning

Many claims have been made about the potential of artificial intelligence and machine learning – with predictions ranging from dystopian wastelands to intelligent societies in which we can all enjoy better health and leisure time. Which is closer to the truth?

Whichever end of the utopia/dystopia scale you tend to err towards, it is worth considering the words of the late, great Stephen Hawkins. Despite benefiting from machine intelligence that enabled him to talk, Hawkins repeatedly warned about artificial intelligence one day spelling “the end of the human race”.

Since the 1980s, advanced computer-based speech prediction helped Hawkins cope with neurological disease. Today, this technology continues to evolve and find wide application in our society.

The same strand of machine learning has helped to power the rise of voice search and the role chatbots could play in the future of elearning.

The potential of AI and ML is vast: technology companies are investing heavily in self-teaching technologies. AI has moved a long way from the “brute force” logic of the Deep Blue computer that defeated world chess champion Garry Kasparov in 1997.

By 2011, IBM’s cognitive computing engine, Watson, won against champion players of the TV show Jeopardy, a language-based, creative-thinking game. Then, in 2016, Deep Mind’s AlphaGo, demonstrated how machines could use neural networks to study a game and learn as it played – by beating world champion Lee Sedol over five matches.

Deep Mind’s latest breakthrough takes the technology one step further. With no prior knowledge or data input beyond the rules of the game, AlphaZero taught itself to play Go, chess and shogi to a level that defeated all human players.

However, while the technology has advanced, the debate has not.

This is what prompted Hawkins to call for greater research into the topic and to warn against the headlong rush to commercialise the technology without fully understanding the potential consequences.

He emphasised the need to think about and plan for superhuman AI – the point at which AI systems go beyond human intelligence processes. While that may be decades away, the rules we programme now into our machine intelligence will be the rules that guide this machine-driven expansion of intelligence.

Hawkins was warning that the mistakes we make now might not be understood until it is too late to put them right; hence his passionate call for greater research, knowledge sharing, understanding and responsible policy making.

Find out more about Stephen Hawkins’ life: https://theconversation.com/a-timeline-of-stephen-hawkings-remarkable-life-93364