The AI fear syndrome is overblown

FEAR about the ill effects of AI on society is not new. In 1968, Stanley Kubrick directed the film 2001: A Space Odyssey based on the ideas of Arthur C Clarke, the brilliant futurist and science fiction writer who also conceived of the simultaneous communications satellite in use today.

The film begins with a sequence about the dawn of man where prehistoric humans discover a mysterious black monolith which triggers a big leap in human cognitive abilities. Then, in a new era, the Discovery One spacecraft is on its way to planet Jupiter with scientists in hibernation to survive the long trip. During this time the ship is controlled by HAL 9000 (Heuristically Programmed Algorithmic Computer), an advanced AI computer system.

HAL can speak, recognise faces, read lips and has other advanced capabilities. It is responsible for all the ship’s functions and communications. When HAL begins to malfunction, Bowman, one of three scientists on board who is awake, decides to disconnect HAL’s high function to prevent further errors. HAL sees that as a threat to the mission and begins to eliminate the crew members by cutting off their life support functions.

Bowman manages to survive and succeeds in disconnecting HAL even while HAL tries to dissuade Bowman and expresses fear and confusion. Of course, the film raises questions about the ethics and reliability of AI. But the important point is that HAL, despite its seemingly “human” qualities, could be shut down – which is precisely what can be done in this day and age.

‘More dangerous than nukes’

A number of prominent people have warned about the dangers of AI. Elon Musk of Tesla and Space X fame has warned about the potential dangers of AI, saying that it could pose an existential threat to humanity. He has called for proactive regulation to ensure the safe development of AI technologies. He is quoted as saying: “We need to be very careful with AI. Potentially more dangerous than nukes.”

Theoretical physicist Stephen Hawking, before his death in 2018, warned that AI could eventually surpass human intelligence, which could lead to unforeseen consequences. He said: “The development of full AI could spell the end of the human race.” Bill Gates, co-founder of Microsoft, said that AI would potentially become uncontrollable.

Jaron Lanier, the computer scientist and virtual reality pioneer, criticised AI’s role in centralising power as well as its effects on employment and personal privacy. He is quoted as saying: “The ideology of AI is that humanity is obsolete, and it’s deeply intertwined with the economic philosophy that digital networks are destined to centralise wealth and power.”

Shoshana Zuboff, Professor Emeritus at Harvard Business School, has talked about the dangers of surveillance capitalism where AI and big data are used to not only predict human behaviour, but also modify that behaviour for profit. He said: “Surveillance capitalism unilaterally claims human experience as free raw material for translation into behavioural data.”

Why then should we consider fears about AI overblown? First, at least at the current time, AI is limited and performs predefined and narrow functions. It struggles with any task that requires very complex decision-making beyond its training data and thus is not creative and capable of adapting to humans. More importantly, as seen in the case of HAL, AI is controlled by humans and depends on human inputs to survive.

We must remember that AI has the potential to bring significant benefits to many fields including healthcare, education and boosting productivity generally. Of course, any technological development is almost always met with scepticism and fear, but history shows that those developments have more often resulted in benefits to mankind. Even in the case of atomic energy where fears of the destructive power of the atomic bomb were well-founded, the benefits of nuclear electric energy are now evident.

In other words, any technology can be used for good or bad purposes. Humans creating the technology are responsible for the beneficial results.

Misguided government intervention

The danger we now face is the prospect of misguided government intervention and regulations which attempt to save us all from the perceived dangers of AI. We are already seeing the tremendous beneficial aspects of AI; generative AI can undertake a wide variety of tasks such as designing new medicines and providing educational programs. However, alarmingly, at the latest count, over 600 AI Bills have been proposed by state legislators in the US this year.

For example, a new Bill in California (SB 1047) would stifle AI development and kill entrepreneurial activities around AI. That California Bill would require AI developers to certify that their AI programs cannot be used to “cause harm”, which is, of course, an unattainable requirement. Since open-source AI models enable the public to access and build their own source code, the developers would then be responsible for third-party activities. Just as a car manufacturer cannot guarantee that an automobile will not kill someone, no AI developer can guarantee their AI model could never cause harm.

The Bill’s contents would be particularly harmful to technology startups because of the high liability conditions. AI developers may move out of California or even out of the US as other states start passing such laws.

California is not the only place where regulation is being considered or even already implemented. In the European Union, draft regulations would create a comprehensive legal framework for AI by categorising AI systems into risk levels – unacceptable, high-risk, limited risk, and minimal risk. Then legal restrictions would be applied.

The Organization for Economic Co-operation and Development (OECD) has established AI principles that focus on inclusive growth, human-centred values, transparency, security and accountability. In addition, a number of countries have imposed bans on the use of facial recognition technology in public spaces as a way to address privacy and civil liberties concerns.

The bottom line is that countries and organisations around the world are focused on the perceived ill effects of AI, and, despite limited knowledge and research regarding the real impact of AI, are coming up with hastily created restrictions and regulations that promise to stifle the growth of this innovative technology.

The answer to this dire situation would be for lawmakers to consult with AI technology experts to learn more about the technology, its strengths and weaknesses, so as to come up with legislation (if needed) which does not stifle creativity and technology generally, but provides a framework to encourage innovation and transparency.

The writer is an investment professional with over 30 years of experience and is co-founder of Mobius Capital Partners