In September 2023, the United Arab Emirates shocked the artificial intelligence world. A government-funded AI lab released Falcon 180b, an open-source large language model (LLM) that, by some measures, was more powerful than the open-source LLMs developed by Google and Meta. 

Overnight, Falcon 180b put the UAE on the tech map; no longer could the desert petrostate be considered a technological backwater. It quickly became clear the UAE was at the forefront of the AI revolution. 

With the benefit of hindsight, the UAE’s strength in AI isn’t a surprise. The UAE government invested in AI early. In 2017, it appointed a Minister for AI. While the British government talks about the need for a comprehensive national AI strategy, the UAE already has one. Effective government policy has helped the UAE give AI heavyweights like the US and China a run for their money. The UK should take note. 

The most obvious explanation for the UAE’s AI success is the sheer amount of money the Emirati government has invested in the sector. The oil-rich UAE has deep-pocketed sovereign wealth funds willing to invest in top-tier computer hardware to train AI models. As LLMs become more computationally intensive, AI companies require more costly hardware to remain at the cutting edge. In March, Abu Dhabi launched MGX, a new AI investment firm with a target of $100bn in assets under management. With the chip shortage and higher interest rates continuing for the foreseeable future, these large cash reserves give UAE-based AI companies a significant advantage.

Additionally, the UAE boasts abundant cheap energy, which makes running these computers more economical than in Europe and Asia. Zero income tax also gives financial incentive for the world’s top AI researchers to relocate to the Gulf state. 

The UAE’s AI advantage is not purely financial. It is also due to astute government regulatory policy. While the cash-strapped UK government can’t match the UAE’s financial firepower, it can create an equally competitive national ecosystem. 

The Sunak government has sought to position the UK as the global leader in AI regulation. There has been some success, with the UK hosting the inaugural AI Safety Summit last November and launching the AI Safety Institute. But the UK can only be regarded as a “thought leader” on AI regulation, because when it comes to legislating, the European Union is far ahead. 

In March, the EU passed the AI Act, the world’s first comprehensive legislation on AI. It is one of the strictest regimes regulating AI and its applications. The EU will now grade all AI use cases into four grades, with the most risky (e.g. facial recognition in public) facing bans.

The sector has broadly opposed the new regulatory regime, arguing it hinders innovation, creates large regulatory burdens and will struggle to cope with future technological developments. 

In contrast to the EU, the UAE has taken a light-touch approach to regulation. The UAE touts the freedoms it offers developers when trying to lobby European AI companies to relocate to the Gulf. In February, the UAE was described by Sam Altman, CEO of OpenAI, as a potential “regulatory sandbox” due to its flexible regulatory regime. 

The UK is currently positioning itself between these two regulatory extremes, opting to wait to bring forward any bespoke AI legislation, and instead relying on existing industry regulators to develop their own guidelines, with some central oversight from DSIT. A White Paper response is upcoming, meaning a bill will likely be put forward in the early stages of the next parliament. This means the UK faces a critical choice, whether to follow the EU or UAE.

With the EU’s new regulation likely hamstringing the continent’s AI sector, many view AI as the UK’s first significant “Brexit opportunity” where a more flexible regulatory regime can attract investment in the UK and drive innovation. 

Questions exist over the UK’s actual ability to diverge from EU policy if market access for UK companies is to be retained. Commentators are drawing parallels with the EU’s 2016 General Data Protection Regulation (GDPR). The so-called “Brussels Effect” occurred as the EU GDPR regime had a significant impact outside of the EU. Tech companies complied globally for simplicity and other governments used GDPR as a template for their own legislation. Many think the same will occur with AI. 

However, AI is different. Variable privacy policies are difficult to maintain on global social networks, but AI companies could easily offer restricted models in the EU while offering cutting-edge, experimental models in more laissez-faire markets. Also, with the economic and political rewards of AI leadership being so great, few countries (including the UAE) will slow their R&D because of the more restrictive EU regulatory regime. As the UK government decides to legislate, it must follow the UAE, giving companies space to innovate.

The UAE offers one other advantage over its competitors, UAE AI companies have been permitted to train their models on private government data. 

The development of meaningful AI applications for the public sector has stalled because governments are unwilling to give developers access to comprehensive data on sensitive sectors like healthcare. Attempts to work with the private sector to better leverage NHS data have received public pushback. The British public is understandably concerned about private companies being allowed to utilise government data.  

In the monarchical UAE, public opinion is undeniably less of a concern. But access is still regulated and the government ensures any medical training data used by AI companies is stripped of identifying information. 

The UK should follow the UAE’s lead in allowing a select group of AI companies to access the NHS’s wealth of data to train their models. The only way AI is going to bridge the gap from quirky consumer applications to transformative technology in the health service is by giving models relevant training data. Given the large number of AI applications within healthcare, the NHS has no chance (or funding) to realise the full potential of AI without the private sector. The government just needs to allow the private sector access to innovate.

This is something that must occur across government. The wealth of UK government data dwarfs that of the UAE, meaning the UK would instantly become the go-to destination for companies looking to turn their experimental models into products with public sector-specific capabilities. The government can easily give UK AI companies an advantage for little taxpayer cost. 

It’s important to remember the UK is still an AI leader. It has one of the world’s highest-skilled workforces and deep private capital markets for start-ups that enable the UK to remain competitive with the UAE’s sovereign wealth funds. The UK must not rely too much on these advantages, government policy must be equally competitive, or else the UK risks losing highly mobile AI researchers and companies to more attractive jurisdictions.

Matthew Brooker is a Policy Fellow at The Pinsker Centre, a campus-based foreign policy think tank focusing on international affairs and foreign policy. The views in this article are the author’s own.