Back in 2014, the late Professor Stephen Hawking of University of Cambridge warned that artificial intelligence, or AI, will pose an existential threat to human civilisation. His concern was that AI will become more intelligent than humans and will effectively take control of the world to our detriment, much as humans have taken over the planet and done immense damage to animal life and the ecosystems. There are potential warnings too from the way invaders subjugated indigenous populations in the past that demonstrate how possession of superior knowledge can have negative impacts on those without. The idea of AI gaining supremacy has also found its way into science fiction and Hollywood movies. At the time Stephen Hawking gave his warning AI was in relative infancy compared to today. Even though he was speaking only nine years ago, a lot has happened in the field of AI since then with AI capability moving forward in leaps and bounds.
Over the last year we have seen the roll out of ChatGPT and other so-called generative AI systems that can answer intellectual queries posed to them with amazingly efficient and competent answers, judged by many to be of a level equivalent to the top quartile of humans. Few people can fail to be impressed by what they can do and there will for sure be big impacts on society, especially in business, education and healthcare, to name but a few fields of application. But will AI soon take over the planet and pose a threat to humans? We might well expect experts in the field to give us a clear answer on this one way or the other.
Unfortunately, that is not the case and we now have vastly differing opinions on the matter. Back in March of this year, over 1,000 global scientists, experts and researchers signed a well-publicised open letter calling for a halt on the development of large AI systems for at least six months until the capabilities, risks and dangers have been fully studied and understood. Signatories to the letter included Elon Musk and Steve Wozniak, the co-founder of Apple. Well, here we are four months later and there is no clear evidence of a pause in AI development —in fact quite the contrary as development continues apace. And now, more than 1,300 scientists and technologists have just signed another open letter, this time written by the British Computer Society, calling for “the UK government and industry to recognise AI as a transformational force for good and not an existential threat to humanity”. So, the experts don’t seem to agree which leaves us in a difficult position concerning the future development and application of AI. In my view the truth in all this is more nuanced than that AI is simply “bad” or “good”.
It is undoubtedly true that AI will have enormous benefits in business with many tasks becoming much easier and faster to do. AI will also transform education and lead to faster and more accurate medical diagnoses. But threats still exist. Even without worrying about whether AI systems will become so intelligent that they will literally take control of the world, the impact on jobs and employment will be profound as will the potential for deep fakes and spreading of false news and information to undermine democracy. There will undoubtedly be winners and losers with the widespread roll out of AI in business. Moreover, the genie is truly out of the bottle and there is unlikely to be a halt to the further development of still more powerful AI. The reason for this is very clear. If any company, government, or research institute decides to stop AI development, there will always be a concern that their competitors will get ahead of them. The race to develop more powerful AI is bound to continue and it will be impractical to try to stop it.
The risk therefore of creating a dangerous “superintelligence” that can outsmart humans is still there and will not go away. Governments need to start to regulate AI development to ensure it is controlled and used in ethical ways and ideally there needs to be global international cooperation on this, particularly to stop a runaway arms race in AI controlled weaponry. Sadly, the likelihood of full global cooperation on this is unlikely given current geopolitical tensions. This leaves us with the prospect that AI is likely to be both a boost for business in the short term and a threat to humanity in the longer term. How long it will take before AI becomes a serious threat is difficult to determine, but as with everything IT related timescales are getting shorter. In the meantime, let’s make the most of AI to benefit business and society but continue to be vigilant about risks and work out how to prevent them.
Graeme Wilkinson is Tan Sri Jeffrey Cheah Distinguished Professor at Sunway University and a Fellow of the British Computer Society.
