The European Commission’s new White Paper on artificial intelligence (AI) may be the most ambitious yet realistic government strategy for AI we have seen.
Aimed at fueling development of an AI ecosystem that fosters innovation and growth for small and medium-sized enterprises (SMEs) while building on traditionally strong European sectors – industrial, transport, agriculture and tourism, for example – the strategy suggests concrete governmental and industry measures for creating a “dynamic data-agile economy” without sacrificing the privacy or personal data of consumers and businesses.
That nuance is important, and often lost in global public debates about AI. Cutting-edge AI technology – pursued by Qualcomm through our R&D center in Amsterdam – will allow AI operations to take place on your device, to avoid sending your data to third parties in the cloud and keep it under your control.
Privacy guarantees and construction of a technological and data-driven economy are not in a zero-sum equation. The Commission’s new strategy recognizes that “building an ecosystem of trust is a policy objective in itself, and should give citizens the confidence to take up AI applications and give companies and public organizations the legal certainty to innovate using AI.”
That’s smart. And there’s so much to be gained. Some examples:
Combined with the ubiquitous connectivity and reliability of 5G networks that started rolling out in Europe last year and will expand significantly this year, AI can already give you a smartphone that instantaneously translates what you say or hear from English into German, French or Portuguese, and vice versa. I have seen this technology employed using a phone that contains one of our newest AI-enabled chips, and it’s impressive.
Other 5G and AI use cases already in trial mode are connected cars that see – or sense – around corners. Imagine you’re driving through the city on a dark and rainy night, and you come to an intersection without seeing another car that’s about to enter the intersection from a different street. Using 5G, the cars communicate to each other what’s about to happen, and the AI in the system tells you to brake or automatically engages the brakes on its own because it also senses no car behind you. AI can help save lives.
5G and AI will also allow your wearable medical device to sense when your heartbeat or blood pressure or blood sugar merits an alert sent to your doctor – but that processes all the information on the device so none of your private data itself is transmitted.
At Qualcomm, we agree with the Commission that AI offers Europe a path to becoming a global leader in innovation for both the data economy and its applications. That’s one of the reasons Qualcomm AI Research for more than five years has teamed up with the University of Amsterdam to push the boundaries of what’s possible when you harness the power of machine learning while enhancing privacy and security. We see the QUVA lab established there as an example of the “Innovation Hubs” envisioned by the AI White Paper, whose goals include ensuring that SMEs can access and use AI, which in turn will create jobs and kindle growth across the EU.
Qualcomm is also making significant contributions to two of the key requirements set for AI by the strategy: technical robustness, and protection of privacy and data. This includes our work and investment into neural network inferences performed on mobile devices and breakthroughs in machine learning. With this technology, as a device collects personal data, AI performs on-device training. Think about it: The AI processing takes place in your hand. Traditionally, machine learning required massive computational power and the transmission of huge datasets across the cloud. So by putting the increasingly capable AI on the device, it also conserves battery and improves the efficiency of the network’s use of radio spectrum.
As my colleague, Max Welling, recently said, it is encouraging to see Europe take control of its own digital future. Max knows what he’s talking about. He’s a research chair in Machine Learning at the University of Amsterdam, a vice president of technology at Qualcomm and one of the foremost experts on AI in the world. “There are great opportunities in combining classical engineering solutions with modern data driven machine learning tools,” he recently said, noting that Qualcomm’s commitment to expand its already sizeable AI R&D effort in Europe “is well aligned with the core values laid out in this commission’s whitepaper.”
I am excited to see how this will play out.