Former Regulator Calls for Agile UN Agency to Regulate ‘AI on Steroids’

Recently, Sam Altman, CEO of Open AI (the creator of ChatGPT) and Omar Sultan Al Olama, Minister of State for Artificial Intelligence (AI), Digital Economy and Remote Work Applications, UAE had a 1-2-1 session on how tools like ChatGPT should or would be regulated.

The Minister, drawing from the movie Freaky Friday (where two people switch bodies and live each other’s lives), asked Mr Altman what one regulatory decision he would take if he switched bodies for a day and became Minister of AI for the UAE.

Mr Altman first responded – lamely – about creating more regulatory sandboxes to experiment with such tools, and use the latter lessons to write regulations. He later changed his answer to having an IAEA-type regulatory “global” system – because he argued that what happens with the most powerful of these AI systems have global impacts, requiring “auditing” and “safety” measures.

He proposed and was granted a “one day conference” in the UAE on this “global” regulation of AI issue, to pre-empt countries doing their own things in an uncoordinated fashion. He wants to see “an action plan” with real buy-in across the world, along with national plans too.

This is my take: I am H Sama Nwana, founding partner of Cenerva and a former senior telecoms/media regulator as an Ofcom UK Group Director. I also completed my doctorate in AI in the late 1980s, and my recent 2022 book (Nwana, 2022) opines, amongst other things, on AI and Big Data regulation.

In an opinion piece in September 2023, I offered the view that unleashed AI on steroids is ultimately unregulatable – but it does not mean we should not try. I asked aloud whether we need a United Nations-style regulation of AI?

I concluded my piece by noting that the United Nations is a flawed organisation, but it is still the least worst option to regulate world affairs today as we see today with Ukraine v. Russia, Israel v. Hamas, etc. I noted that a United Nations-mandated agency called the ITU ultimately regulates the world’s radio spectrum resources.

Indeed the most recent four-yearly World Radio Conference (WRC 2023) just concluded in December 2023 in Dubai (UAE). I wondered aloud in my piece how AI on steroids should ultimately be regulated too using another UN agency. Interestingly, Mr Alman mentioned one such UN-agency model – the IAEA.

The IAEA stands for the International Atomic Energy Agency. It styles itself as “the world’s centre for cooperation in the nuclear field and seeks to promote the safe, secure and peaceful use of nuclear technologies”. The IAEA is indeed an autonomous intergovernmental organisation that was established in 1957 within the United Nations system, that seeks to promote the peaceful use of nuclear energy and to inhibit its use for any military purpose, including nuclear weapons. It reports to both the General Assembly and the Security Council of the United Nations.

Therefore, it appears to me that Mr Altman is settling on a similar position to mine, i.e. being regulated through some international UN regulatory agency, albeit modelled around some organisation like the IAEA.

On the one hand, Mr Altman’s mention on the IAEA model makes some sense because AI would be regulated to promote its ‘peaceful’, benevolent use, whilst inhibiting cyber-military or other malevolent uses. This is the theory.

On the other hand, this model makes zero sense because the IAEA is a tortoise-like, slow-moving, deliberate international regulator of a physical and complex-physics technology – nuclear technologies. There are less than a ten nuclear powers in the world today in 2024, including a maverick one like North Korea.

AI is fast moving – literally at the speed of light – and is already in the hands of practically every country, including ‘rogue’ and maverick states.

Mr Altman literally talked about imagining a company with an ‘AI programmer’, an ‘AI lawyer’, an ‘AI marketer’, presumably replacing – if not supplementing – their human counterparts. This is no longer far-fetched vision with the sort of models that Open AI’s ChatGPT mission are evolving – and they are open-sourcing some of the models like ChatGPT 2 is already today.

The future company with an ‘AI programmer’, an ‘AI lawyer’, an ‘AI marketer’, etc. reads very benign and benevolent.

Replace these with an ‘AI bombmaker’, an ‘AI AK47 gun printer’, an ‘AI cyber hacker’, an ‘AI voice cloner’, an ‘AI deep fake expert’, etc. – and where do we find ourselves?

Rogue companies and countries would set up organisations to do these too. The challenge is that these ChatGPT-like AI models (and more) are in the hands of smart university students and more in all countries across the globe.

I argued in my last opinion piece that trying to regulate the genie-out-of-the-bottle now is ultimately not possible.

That “thoughts” and “thinking” that humans self-regulated in the past have now been unleashed on to digital machines. That “rogue [and non-rogue] countries” interests want to use AI to upend their perceptions of Western dominance. “AI on the loose” and deliberate cyber-warfare from rogue states is a toxic, but it is already here.

I opined that sheer human ingenuity and “thinking” which drives all civilisation is the at the core of a new AI on steroids, which combined with benevolent Big Tech industry interests and non-benevolent rogue state interests too – make it ultimately unregulatable.

However, one thing I know for sure is that we must do our best to regulate AI – and fast. Speed is of the essence. A UN-specialised body should be mandated – like the IAEA was mandated – to start regulating AI.

But that is where the comparison should end, as a tortoise-like, slow-moving, deliberate international regulator of AI would be toothless and rendered obsolete before the ink dries on its set up. The United Nations (UN) already plays a crucial role in shaping international rules and governance for Artificial Intelligence (AI).

So my answer to the Minister’s question to Mr Altman that I used to start this piece is simple – the UN should just get on with it and set up the most agile and fast-moving regulatory agency it has ever set up.

AI is on the loose already – and some of it needs to be reigned in. No time to waste. I am now off the fence from my September 2023 piece.

We need the United Nations AI Regulator – now!

I realise it is not something the very smart Minister of AI for the UAE can do on his own, but he can definitively push for it. The subject of his next one-day conference in the UAE should be on the “action plan” to create the UN agency for regulating AI – and one with teeth and tools to do the job.

The conference should be the start for the “action plan” for the inevitable UN AI regulator with all countries invited to provide real “buy-in” from across the world, as Mr Altman seeks.

About the Author

Prof H Sama Nwana, founding partner of Cenerva and a former senior telecoms/media regulator as an Ofcom UK Group Director. He completed his doctorate in AI in the late 1980s, and authored a recent book – The Internet Value Chain and The Digital Economy (Nwana, 2022) which opines, amongst other things, on AI and Big Data regulation.