On March 13, the European Union’s parliament passed the historic EU AI Act, the first major act crafted to regulate AI or Artificial Intelligence. The set of laws, the result of almost two years of work, covers various AI applications, from spam filters to biometric identification systems.

Why regulate AI?

AI’s swift advancement and proliferation across different industries have led to rapid development, from enhancing efficiency to sparking innovation. However, the quick changes also spurred concerns about the potential negative impact on society, the economy, and individual rights. The more AI systems become integrated into various facets of human life, the more inquiries and apprehensions regarding their potential risks and responsible deployment emerge.

Categorizing AI systems based on risk

One of the critical components of the act is how it will classify AI systems into different risk categories. The classification will be based on their potential impact on safety, rights, and societal values. The higher the risk classification, the stricter the rules.

However, AI systems that fall under the “Unacceptable Risk” classification could be banned altogether—barring very specific exceptions, including law enforcement purposes. Tools that pose significant threat such as deepfake technologies that can be used for deceptive or fraudulent purposes. Another example are autonomous weapons which can make decisions that pose serious risks to human life and international security.

Legislators believe the monumental act’s risk-based approach will help ensure check and balance for the industry players and developers. It is a big step towards harnessing this relatively new technology more responsibly and ethically.

Criticism from stakeholders

Like any regulation, this act has also received its fair share of criticism from various stakeholders. An open letter signed by tech founders, startups, and even well-established companies like carmaker Renault summarizes the sentiments shared by industry players.

First, there is a question about how the law will stifle innovation and the development of Artificial Intelligence. This, in turn, can lead to unwanted results, including pushing AI companies to limit deployments of their tech in the area. And when AI developers choose to operate in less regulated markets, tech development in the EU will surely suffer and European technology will start to fall behind.

Aside from the noted ambiguity and complex categorization of the AI systems, another issue argued in the letter is the burdensome cost of complying with the act itself. This can disproportionately affect smaller businesses—SMEs and startups—and significantly impact their ability to compete with larger companies.

Looking ahead

After final formalities and the go-ahead from EU member countries, the world’s first comprehensive AI law will officially be enforced. According to reports, the target period for enforcement is May or June of this year. It’s important for business that’s operating in the EU or marketing towards

Whether or not its enforcement will be a success is still up in the air. However, it’s clear that this is just the start of laws regulating Artificial Intelligence. This landmark act will surely influence the rest of the world. Other countries and governments are taking notes on how this act will fare and how they can find ways to make AI systems work better for the people who use them, and not just the companies or institutions who deploy them.

Striking a balance

It’s true that there is a need to create guardrails for this tech to protect the rights and safety of all. But it’s also important for lawmakers to strike an elusive yet critical balance between setting rules and supporting growth and innovation. One way that lawmakers can do this is by working closely with industry experts who are willing to share their expertise and cooperate in order to make the future of AI better and safer for everyone involved.