There is no AI – race and if there is, it’s the wrong one to run
By Virginia Dignum – co-founder ALLAI
Press and policy makers are obsessed with the so-called AI race, and with Europe’s position in it. Just this week at Davos, US executives warned that China may be winning this supposed race. In another recent article, Bloomberg pointed out that countries are rushing to not be left behind. It also correctly pointed out that there is still a long way to go before Artificial Intelligence (AI) will be commercially viable. In its vision for AI, launched last December, the European Commission has described its concerns with the position of AI in this race, which some have claimed Europe already lost.
In my opinion, this race discourse is both wrong and dangerous. It puts the focus on competition and brings with it a sense of gloom and despair. But let me share two insights with you.
Firstly, there is no race
and secondly, if there is, it is the wrong race to run.
There is no race because of the very definition of a race: a competition of speed, against an objective criterion, usually a clock or to a specific point. In AI developments, we don’t have a end point! Nor do we have a specific time to stop. There is therefore no way to determine when and where someone will win this so-called race. By assuming that this ‘race’ can be won, means that there would be a moment in which we can stop developing technology, and advancing humankind. There may be some battles to be won, but winning one battle does not mean that the victor is ready.
That it is the wrong race to run, is an even more important issue. The US and China are betting on machine learning developments, and in particular on deep learning, as the approaches that will achieve AI, and thus enable them to ‘win’ that so-called race. These approaches rely on the availability of huge amounts of data and computational power, to enable machines to perceive, or learn, characteristics of a particular domain. This is used to recognize faces in pictures, to determine the credit worthiness of a mortgage applicants, or to recognize cancer cells in scans or X-ray images. All of it relevant and important applications, and the progress achieved in the last few years is truly remarkable. However, these approaches are focussing on one aspect of intelligence, that of the ability of perceiving patterns and make predictions based on those perceptions. True intelligence is more than this, and includes the capabilies to reason, interact and and decide based on little, incomplete and contradictory information. There is an urgent need to explore alternatives for statistical approaches to learning. In fact just this week, a study analysing 25 years of AI research has concluded that the era of deep learning is coming to an end. Europe has tradiotionally always been strong on symbolic approaches to AI and on (social) robotics. These are exactly some of the areas to be investing on at the moment, the technologies that will bring AI forward in the near future. In would therefore be a mistake to blindly follow US and China on their machine learning ‘race’ when now is the opportunity to show the worth of alternatives approaches. Approaches in which we Europeans may have an advantage.
Other reasons by which focusing on data heavy approaches is not the thing to are their negative impact on human wellbeing and on the environment. Any development that does not boost trustworthiness will ultimately not succeed. There is no business model for untrustworthy AI or unethical AI. The results and decisions taken by systems based on deep learning and neural networks are hard to understand and to explain and therefore not sustainable in many areas in which the trust of users and experst is crucial. Moreover, current approaches are extremely environment unfriendly: the amount of (energy) resources needed to store and compute data are already comparable to the needs of a small city. This is not sustainable especially if the way forward is based on the exponential growth of data and computational power.
Europe is home to strong, world leading, fundamental research in AI and is known for a strong ethical background and respect for human rights. Putting these at the core of advances in AI will lead to great breakthroughs that can truly bring AI forward is ways that are both financially profitable but foremost, promoting of human and environment well-being. This will imply a mind shift when it comes to how we do business and how inclusive the decision-making process is. Developing AI responsibily, grounded on ethical principles and human rights, is not a burden on research and investment but THE stepping stone that will bring this powerful technology forward. More than a technical decision, this is one of policy and vision which only Europe is able to realise at this moment.
The aim is not to win races, the aim is to ensure wellbeing of humankind and environment.