Dr Jannie Zaaiman, CEO of the South Africa Information and Communication Technology Association.
Artificial intelligence (AI) has both the potential to drastically change the world for good, and the ability to wreak havoc like a dystopian novel in which man meets machine. It is, with apologies to Charles Dickens, both the best of times and the worst of times.
Harnessed correctly, AI and its contribution to the fifth industrial revolution can be almost immeasurable. Instead of replacing people, this cutting-edge technology can serve as a cognitive partner that augments human judgement, creativity and productivity.
Industry 5.0 promises harmonious machine and human collaboration for the benefit of all. One of these promises should result in humans and machines collaborating in ways that bring more sustainable and ethical practices to bear.
However, without meaningful oversight, this technology has the potential to run rampant through our lives, leaving a trail of destruction in its wake. Then we’ll be fighting AI robots in a war like never before – WWIII even, except the enemy is far more powerful than ever before.
Any legislative process will always be playing catch up.
The current situation is that there really is no actual oversight. Yes, there are movements towards a regulatory environment but that’s all they are in an environment where technology is light years ahead of any regulatory system.
Any legislative process will always be playing catch up.
In Europe, there’s the AI Act, which the European Commission says “is the first-ever legal framework on AI, which addresses the risks of AI and positions Europe to play a leading role globally”. And this is a global first, it says.
Dated 2024, the concept is that AI should be developed and used responsibly within a set of clear and risk-based rules for anyone building or deploying AI systems. And, as the EU moves to this new regulatory framework, the European Commission has introduced the voluntary AI Pact, which is a sort of ahead-of-time compliance commitment by players in the sector.
Further back in time, we have the OECD AI Principles, which the 38-member country organisation claims is the “first intergovernmental standard on AI”. These principles, says the Organisation for Economic Co-operation and Development, “promote innovative, trustworthy AI that respects human rights and democratic values”.
They were adopted in 2019 and updated in 2024 and comprise five values-based principles and five recommendations that provide practical and flexible guidance for policymakers and AI actors. Members and non-members alike have pledged to respect the rule of law, be transparent, ensure robust security and be accountable when it comes to implementing AI.
Across the Atlantic, the United States has relied heavily on voluntary commitments from the likes of Amazon.com, Google, Meta, Microsoft and OpenAI, which MIT Technology Review points out hasn’t gone far. The then US president Joe Biden went so far as to invite heads of these companies to the White House. What that country seems to have instead, are laws at state level.
And regulation is something that people want to see happen. Earlier this year, KPMG in association with the University of Melbourne, published a global study into trust, attitudes and use of AI. This paper is based on the views of more than 48 000 people from 47 countries across all global geographic regions. Some 70% of those surveyed believe regulation is necessary, yet less than half (43%) think that current laws are adequate.
There’s a list of what the respondents want that includes international laws, national government regulation, co-regulation with industry, and legal means to fact-check AI-generated misinformation.
Self-regulation doesn’t seem to be much of an option given that Forbes has pointed out that there has already been what it calls a “string of AI fails”. There’s an AI incident database that has catalogued more than 1 200 issues when AI failed or was misused.
The only logical conclusion we can draw from all of this is that there may well be laws, regulations and oversight mechanisms but these are not only outdated, but also too incongruent to have any real effect on a technology that will always move faster than anyone can put pen to paper.
What we really need is a set of global guardrails – a living framework – set up at World Economic Forum level that regulates AI globally and is flexible enough to keep pace with AI’s evolution while setting boundaries on how it may be developed and deployed. Otherwise, we have a serious problem.
