phonlamaiphoto - stock.adobe.com
UK and EU AI communities will only be globally significant if they work as one
The European Union and the UK will be left behind in artificial intelligence research and development if they take separate paths after Brexit
The US and China will dominate the development of artificial intelligence (AI) unless the European Union (EU) and the UK can find a way to continue to work as one after Brexit, according to one of the founding academics of a European collaboration project.
Both communities are too small to generate the capital and skills needed to compete with the US and China and could miss an opportunity to take AI in a different, more balanced, direction, said Holger Hoos, professor in machine learning at Leiden University in the Netherlands.
“If you were to take the UK community out of the EU’s AI community, you would be left with two pieces, neither of which would have critical mass,” he said.
Hoos is one of three founders of the Confederation of Laboratories for Artificial Intelligence Research in Europe (CLAIRE). Its biggest challenge, he said, is “to make sure that European AI efforts reach critical mass”, adding: “This is rather than diluting everything to the point where the enormous amount of money that is going to be invested doesn’t have much of an effect.”
Asked whether the UK AI community will play a role in the organisation after Brexit, Hoos said: “You bet, at least if we have anything to say about it.
“If you look at where AI talent and capability sits in Europe the UK has been one of the most important places since the beginning. But as nice and important as the UK is, I think the AI community there is too small to compete and the EU community without the UK also doesn’t have critical mass.” From CLAIRE’s point of view, that would be a complete disaster, he said.
The organisation has hundreds of UK members, including academics from the University of Oxford and Imperial College.
Hoos said the project, although mainly funded by the EU, would be based on geographical Europe rather than just the EU. “Eighteen months ago, we saw various paths through Brexit and from the beginning, it was very important to conceive something that would draw from and support Europe as a geographic region,” he said.
“We made it very clear that no matter what happens with Brexit and the relationship between the EU and the UK, in AI we should work together.”
The EU and the UK will have to work closely together if they are to have any significance in global AI developments, said Hoos. The US and China are the biggest players, but there is an opportunity for another region to set a new path, he added. “China and the US are very different to each other and different to Europe in how AI is seen going forward.”
Hoos said that in the US, a lot of AI research is done by private companies such as Google and Facebook, which gives it a particular flavour. “There is nothing wrong with this, but it is biased towards business and, to a lesser degree, consumer interests,” he said.
Read more about artificial intelligence
- A Tech Nation programme to support the UK’s financial technology startups demonstrates the increasingly diverse range of business-to-business products and services available through the country’s fintech community.
- The $1bn multi-year investment should put Microsoft developers at the forefront of coming up with artificial intelligence tools for the Azure Cloud.
- Alphabet’s DeepMind company loses millions of dollars each year, but it appears the research firm’s AI advances are so valuable that Alphabet finds it worthwhile anyway.
In China, AI research is very much state-controlled, said Hoos. “So it is biased towards what they consider to be the most useful things for society as a whole,” he said.
“We think that in Europe, we should find a balance. This is not just between two poles of society versus private enterprises, but also including the needs of private citizens.”
Hoos said CLAIRE wants a brand of AI in Europe that is also pointed firmly at the common good, such as helping to meet sustainability goals or introducing AI that works for non-government organisations for the benefit of citizens.
“We want to help ensure that Europe can play a leading role in AI research and innovation,” he said. “We believe, as many do, that AI is a very important set of technologies going forward, not only for economic prosperity, but for the benefit of society. We therefore think it is very important that Europe as a geographic region is doing the right thing with AI in order to remain competitive and do the right thing for European citizens.”
Europe’s inherent advantages and focus will help it achieve this, said Hoos. “Europe has the opportunity if it starts out on this path and it also has the advantage of having a diversity of perspective and talent,” he said. “If you find something that works for all the slightly different interests and cultures in Europe, it is very likely to work far beyond as well.”
CLAIRE was set up in June 2018 with an open letter from 600 of Europe’s best AI scientists to rally more people to the cause. Its members were initially just academics, but then grew to 3,000 including people working in industry as well as those in politics. “But the majority of people basically have a PhD in AI,” said Hoos.
So far, the organisation, which has a number of small offices across Europe, has been funded by its members and the academic institutions that support it. Ultimately, the organisation hopes to receive funding from the EU and its member states as well as associated states, such as the UK after Brexit.
Central headquarters
It recently opened its headquarters in The Hague because of its central European location and the Netherlands’ strength in AI research. “The city also has a reputation for peace and justice, which relates to our view of how AI should be taken forward for the benefit of society,” said Hoos.
CLAIRE’s most advanced work to date is being done in collaboration with the European Space Agency (ESA). CLAIRE has been in existence for only 18 months, but this work began about a year ago. It is working with ESA to put critical skillsets into Earth observation and other areas in which it is active.
“One of the areas this makes a big difference in is work on climate change, which is dependent on Earth observation,” said Hoos.
Currently, the organisation’s main aim is to ensure that money being invested in AI by European governments is effective and that investments help keep AI experts in Europe, despite the attraction of moving to the US to progress their careers.
When it comes to investment in Europe, Hoos said the EU’s plan to eventually spend €20bn a year is “barely enough”, although it is of the right magnitude. Just the academic curiosity-driven research part of AI exceeds €1bn a year, he said.
Hoos said that beyond China and the US, there are small countries, such as Singapore and the United Arab Emirates (UAE), that are making leaps in AI. “Europe risks being left behind if investments are not effective,” he warned.
UAE’s growing challenge
For example, in the UAE, a university focused specifically on AI has been set up in Abu Dhabi. Mohamed bin Zayed University of Artificial Intelligence (MBZUAI) has been established to ensure the Gulf state keeps pace with rapid global developments in AI, which will be essential as it diversifies its economy away from a heavy reliance on oil and gas.
Oxford University professor Michael Brady is interim president of MBZUAI, and he told Computer Weekly in November that the UAE government’s AI strategy puts UK government planning to shame. “There is a stark difference between the short-termism that characterises so much of government policy in the UK, where politicians worry about the headlines tomorrow morning,” he said. “It is so refreshing to be part of a government-led initiative that has a 30-year vision to transform the economy and the culture.”
A key part of CLAIRE’s plan is to create a large AI hub in Europe, said Hoos. “We do not think that all AI resources need to be centralised in one place, but we do think there should be one place where Europe’s top scientists come together, do large projects, talk to each other, exchange best practices, and so on.”
The AI hub plan will take a long time to achieve, he said, but CLAIRE members are already talking to MEPs to get them interested in the idea.
“The location for the hub should be one committed to the EU as this will be where much of the funding comes from,” said Hoos. “It would have to be central, and it would need a lot of local talent and should be a place that is practical to live in.”
