The European Union may struggle to set global standards for AI development and usage because of its lack of leading-edge AI research firms, says a report.
Businesses in the European Union are at risk of becoming too reliant on artificial intelligence technologies from China and the United States, which are responsible for almost all of the leading-edge AI operations.
That according to a report by the Future of Life Institute, which claims that the EU has not been successful in fostering a “general purpose AI system”, by which they mean a system such as OpenAI’s GPT-3, Google’s LaMDA, and DeepMind’s Gopher.
SEE ALSO: Europe’s Digital Decade Requires Collaboration and Data
“European companies are not developing general-purpose AI systems and are unlikely to start doing so anytime soon due to their relative competitive disadvantage vis-a-vis American and Chinese players, and due to the amount of money, data, and computational resources it requires to develop these systems,” said policy researcher at the Future of Life Institute, Risto Uuk. “Instead, European companies will likely rely on systems that are developed elsewhere.”
Not only could this affect European businesses ability to develop their own unique AI models, as they will be based on the foundation model from an American or Chinese firm, but it could lead to the European Union struggling to create global standards for AI.
Uuk sees a scenario where most European businesses will utilize APIs provided by larger American and Chinese firms, such as Baidu, Google, Meta Platforms, and OpenAI. As we move from research to monetization, these APIs could become a costly expense for European businesses.
Many European businesses are already utilizing AI models built by OpenAI and others, signifying the gap between researchers in both regions. It also shows the inability for leading-edge AI research to be conducted without access or a partnership with one of these well-funded firms.
The European Union published its draft legislation for AI regulations in September, which were meant to give AI developers plenty of time to look at the regulations and see what is required. The EU is likely to push forward with the regulations sometime in late 2023.
