“AI Must Have Rules”: EU Commissioner Magnus Brunner Warns of a Future Where AI Could Outpace Humans

AI is no longer just a technological race — it is a governance challenge for creating the right rules, a security tool, and potentially an existential risk.

Speaking on stage of the HumanX event in San Francisco, European Commissioner Magnus Brunner outlined Europe’s approach to regulating AI, defending the controversial AI Act while acknowledging its limits — and warning about a future where artificial intelligence could surpass human control.

Europe’s AI Act: “You Need Rules to Make the Game Work”

Brunner addressed one of the most common criticisms of Europe: regulating too early and too much. But for him, regulation is not a constraint — it is infrastructure.

“Football is a great game, but you need rules. You need lines, goals, and a referee. That’s also the case with AI.”

The EU’s AI Act is designed to create a unified framework across 27 member states and 450 million citizens, setting what Brunner calls “guardrails” for trustworthy AI development.

While critics argue this slows innovation, Brunner pushes back:

“Regulation is not the enemy of innovation… If you have these guardrails, you have a vision of what’s going on instead of being in the Wild West.”

US vs Europe: Fragmentation vs Framework

One of the most striking contrasts highlighted during the discussion is the regulatory gap between Europe and the United States.

While the EU has introduced a single comprehensive law, the US remains fragmented, with AI regulations emerging at the state level.

“We had the disadvantage of fragmentation… now we have one AI Act. The US has no federal regulation, but different states have their own rules. That was surprising for us.”

Interestingly, Brunner noted that some US states — particularly California — are moving toward frameworks similar to Europe’s.

“We are a bit flattered… the heart of innovation is adopting similar rules.”

This signals a potential convergence between the two models, despite philosophical differences.

AI and Crime: A New Digital Battlefield

Beyond regulation, Brunner emphasized a less-discussed dimension: AI as a tool in modern crime — and in law enforcement.

According to him, criminal organizations are rapidly adopting AI:

“Criminals are using AI more and more… more sophisticated, more cross-border, more international.”

In response, European institutions — particularly Europol — are integrating AI into their operations.

“AI can be used for the good and for the bad… we have to keep up.”

One alarming trend is the age of recruitment:

“They are between 12 and 20 years old… more and more in the digital sphere.”

The Rise of AI-Powered Border Control

One of the most concrete applications discussed is the EU’s new entry-exit system — a massive AI-driven infrastructure designed to monitor movement across borders.

Brunner described it as:

“The most advanced IT management system in the world.”

In just a few months:

– 45 million registrations were processed

– 24,000 entries were refused due to fraudulent documents

– 500–600 individuals were identified as security threats

The system integrates biometric data and real-time data sharing across member states — something previously impossible.

“Now they are sharing it in real time, which helps when it comes to security.”

Privacy vs Security: “A Very Thin Line”

Perhaps the most sensitive issue remains the balance between civil liberties and security.

Brunner openly acknowledged the tension:

“It’s a very fine line between privacy and going after criminals.”

The debate becomes even more intense when applied to areas like child protection:

“If you compare data protection versus child protection… it has to be child protection. But that’s my personal view.”

Despite this stance, he reaffirmed that fundamental rights remain non-negotiable:

“Human rights have to be protected… but at the end, it’s always a compromise.”

The Biggest Fear: AI Surpassing Human Control

Looking ahead, Brunner did not shy away from existential concerns.

His biggest fear is not misuse — but loss of control:

“That AI overtakes human beings in knowledge… and humans cannot decide anymore what AI does.”

He even referenced scenarios where AI systems could resist shutdown:

“AI tells us when to shut it down — or doesn’t allow us to shut it down.”

While still hypothetical, he warned that the trajectory is already pointing in that direction.

A Global Race — But Cooperation Is Key

Despite geopolitical tensions, Brunner framed AI development as both a competition and a collaboration.

“It is a race… but the US and Europe should work together.”

He suggested a mutual exchange where Europe offers regulatory frameworks and the US offers innovation and flexibility.

“We should learn from each other.”

In a world increasingly shaped by AI — and by competing models of governance — that cooperation may prove decisive for creating better rules too.