The European Commission has announced that it will co-fund four testing and experimentation facilities, or TEFs, for AI exploration, in order to support responsible AI development and ease compliance with the EU AI Act, which is in the final stages of approval before being enacted into law.

The TEFs will function in both the physical and virtual realms, according to a statement from the Commission this week. Work at the four initial TEFs, which began at the beginning of 2023, centers on agriculture and food production, healthcare data, smart manufacturing and smart city technology.

Funding for the TEFs will total between $43 billion and $65 billion, which will come from both the Commission itself and from member state contributions. The TEFs will run, at least initially, for five years.

The smart city TEF is known as CitCom.ai. A statement published by senior researcher Martin Brynskov of the Technical University of Denmark — the project’s host — describes it as focused on AI solutions that could be applied to urban transportation, including self-driving cars.

“With an initial focus on power, mobility and connectivity, its job is to test AI rand robotics before they get into places where humans live and move around,” he wrote.

AI-Matters is the manufacturing-focused TEF. It appears to be in an earlier stage of development than CitCom, but the project’s website said that it will be focused on services like factory optimization, human-robot interaction and the adoption of new AI technologies. A catalog of those services will be available by September of this year.

AgrifoodTEF will focus on several distinct areas of agricultural technology, including autonomous weeding and arable farming techniques, AI optimization of growing conditions for tree crops and horticulture, AI management of livestock operations and feed production for increased sustainability, and data modeling and standardization for the food processing industry.

TEF-Health will be made up of hospitals as live testing sites as well as laboratory testing facilities for AI and robotics in the healthcare sector. The use-cases, according to the Commission, will focus on cancer care, neurotec, cardiovascular health and intensive care.

Testing and validation via TEFs is expected to help AI technology developers comply with the AI Act, which passed the European Parliament earlier this month. That law includes provisions regulating several aspects of AI systems, including transparency requirements for AI-generated content, a ban on AI biometric surveillance systems in public spaces and stiff testing requirements to mitigate “reasonably foreseeable” risks to society.