EU AI Act, a chance or risk for the technological innovation in europa

This is image is not from me and might be subject to copyright

Europe is laying the foundation for regulating Artificial Intelligence technology with its AI Act.

The AI Act aims to be the world’s first comprehensive AI law, addressing a society where AI based products are increasingly adopted without undergoing the testing phase typical of most industries. However, since the approval of the AI act draft texts by the European Parliament last June, many voices have arisen against its content, claiming that it will jeopardise the innovation in Europe. In fact, The law categorise all AI systems in four groups:

• prohibited ones such as government social scoring and biometric identifications systems in public space
• high risks ones such as autonomous driving systems
• limited risks systems such as chatbots
• minimal risks systems such as spams filters

When being available in EU, these systems will need to comply with law-based requirements that vary in strictness based on their categorization. If successful enacted, Europa will pioneer AI regulation, potentially influencing other countries to adopt similar laws. US Senat, some regulators authorities in the US, and other countries like Brazil closely monitor the draft law’s progress indicating that concerns about AI regulation extend beyond the EU. Members of the European Parliament favoring AI regulation welcome this law, as it compels companies to be more transparent and accountable for their AI based products.

In the industry, opinions on the law are divided. Tech leaders behind the most prominent AI applications argue that the law will over-regulate the products within the EU making difficult for them to comply with it. On the other hand executive of big companies fear increased costs and an overload of work to assess their current systems and applications and making the AI based ones compliant. In a competitive and rapidly advancing technological landscape, investing time and resources on compliance matters could become a disadvantage, especially if the laws are out of touch with reality.

The law is currently under discussion; the European parliament, council and commission are jointly reviewing it. The law texts are expected to be amended and adopted by the end of this year, becoming official the following summer. A two-year transition period will be given to companies to comply with it.

While the AI act will end a laissez-faire approach for tech enterprises and bring transparency to AI-based tools’ users, AI is and will remain the groundbreaking technology of the past decade; and the US and China have a clear advance over Europe. That is partially due to less bureaucracy and fewer/inexistent laws to comply in their ecosystems. To avoid falling behind and merely becoming consumers of AI-based products, Europe must ensure the law provides a solid, flexible framework to foster AI research and development within its ecosystem, focusing on regulating AI-based products and not the technology itself.