BRUSSELS — The EU says it has a new problem when it comes to artificial intelligence: Companies are not using it.
As Europe struggles to counter America’s overwhelming dominance in the artificial intelligence space, a strategy out Wednesday and seen by POLITICO will target faster adoption of the technology as a way to turn things around.
European companies have been slow to deploy AI to change the way they’re working. Less than 14 percent of European businesses used AI in their activities last year, well behind the global trend.
Lucilla Sioli, the boss of the European Commission’s AI Office, admitted at a POLITICO event this month that’s “not a great number.”
The solution? “Targeted measures to boost AI use in key sectors of the economy,” according to the draft of Wednesday’s strategy. It outlines ways to integrate the technology into 11 industries, ranging from manufacturing and defense to health care and mobility — with proposals including supporting AI models for autonomous driving and drug discovery, as well as an app store for farmers.
As places such as the U.S., the Middle East and China put the technology front and center, Europe’s slow embrace risks squandering the bloc’s best chance to restore its waning economic growth — even as the EU has failed to develop its own models.
While Wednesday’s strategy is not about winning the race for homegrown artificial intelligence models, it acknowledges that efforts to build the underlying technology so far have been unsuccessful. Europe’s “dependencies” on other regions for AI hardware could be weaponized and pose a significant supply chain risk, the draft says.
Yet companies shouldn’t hold back from using the technology, according to the executive.
“Whenever a company or public office faces a new challenge, the first question must be: How can AI help?” Commission President Ursula von der Leyen told an event in Turin on Friday ahead of the release.
Getting down to business
While 4 out of 10 large companies in Europe used some sort of AI technology last year — including text mining, generation and creating images — that figure stood at 1 in 10 for smaller companies.
Although data is patchy, that’s a big difference from estimates elsewhere. The U.S. Chamber of Commerce estimates that almost 60 percent of small businesses in America use AI. A survey by McKinsey reckons the use globally could be as high as 78 percent.
Investing in AI is more affordable for companies that can roll out technology at scale. But companies also have to figure out where to invest, which is seen as a major hurdle.
The Polish government surveyed companies on their willingness to invest in cloud and AI. “More than 600,000 Polish companies want to invest in cloud and AI in the next six to eight months,” Polish Digital Minister Dariusz Standerski said in an interview in September.
He argued that money is not always the problem, and neither is the burden of changing processes to roll out technology. The problem is instead that companies “only see …. universal solutions,” and not a “specific solution suited for [their] company,” Standerski said. For many, AI is primarily still associated with mass-market chatbots, such as OpenAI’s ChatGPT.
That leaves many specialized companies wondering what it can do for them. The draft strategy aims to address this, with the EU set to support a range of improvements for various industries and use cases, and roll them out at scale.
Most are small steps. For example, the EU plans to support a farmer-focused app store to allow farmers to browse and discover AI-powered apps, according to the draft.
For robotics and manufacturing, it plans so-called “acceleration pipelines” to speed up AI-powered robotics and manufacturing solutions. For the creative industries, it would support studios specializing in AI-enhanced production and develop a platform to utilize AI translation, making foreign-language news more widely accessible.
Some of the initiatives are more ambitious in their plans to address dependence on foreign-owned AI, for example, such as a plan to ramp up support for European artificial intelligence models for real-time understanding of the battlefield.
Top Commission officials have raised the bar for themselves ahead of the release, arguing the goal of the strategy is to plug AI into companies’ core activities — not just for support services.
“What we want to see is that companies integrate AI in their production process,” said Sioli of the Commission’s AI Office. “It’s not about having ChatGPT on your desk when you go to work, and then you tick a box that you use AI.”
“We are looking at the core processes, because the supporting processes, that’s relatively easy, already many companies have taken it up,” Małgorzata Nikowska, the AI Office’s head of unit for innovation and policy coordination, said at a hackathon organized by OpenAI and startup lobby group Allied for Startups in mid-September.
The real test of success, she said, will be whether Europe can persuade companies to redesign their production processes to use AI.
The post European businesses barely use AI. Brussels wants to fix that. appeared first on Politico.
