On May 21, US Vice President J.D. Vance sounded the alarm: the USA must not put the brakes on the development of artificial intelligence for security reasons - otherwise there is a risk of dependence on Chinese AI systems. Such "enslavement" by foreign technology is inconceivable for him.
Tech giants such as OpenAI and Microsoft are also warning against too much regulation and calling for more leeway to secure the USA's AI lead - and politicians are following suit.
Donald Trump even negotiated an AI agreement worth billions with the United Arab Emirates, which is intended to further extend the US lead.
More than one trillion dollars have been earmarked for AI data centers by 2030 alone.1
Meanwhile, AI pioneer Yoshua Bengio urges caution: instead of just focusing on winning the AI race, the world needs to start thinking about how we can survive this race safely. According to Bengio, the focus is currently too much on speed - and too little on safety. We are now all spectators in a fast-paced, barely controllable race.2
The dark side of the AI race - risks and dangers
What happens when the will to win is greater than reason? The following three risks show how dangerous an AI arms race can become without clear boundaries and responsibility.
Safety standards? No answer
In the heat of progress, optimizations, publications and improvements are made - but often without thorough testing.
Inadequately tested AI models get into circulation, spread uncontrollably - and the risk of misuse increases with every update.
What if such an AI suddenly launches cyberattacks? Or sends out swarms of drones because it has been trained incorrectly?
An AI arms race may create systems that are too powerful to control, especially if we don't worry about losing control in the first place.
Ethics? A minor matter
When market shares or geopolitical power are at stake, values such as transparency and responsibility quickly fall by the wayside. In authoritarian states, AI has long been part of surveillance: facial recognition, movement profiles, social credit systems.
But even democracies are not immune - they stretch ethical boundaries when it benefits the economy. As a result, we are moving dangerously close to the border between control and loss of control.
Loss of control? More realistic than expected
With the exponential progress of AI, there is an increasing risk that at some point we will develop systems that we can no longer fully understand or control. Even today, many AI models are "black boxes" whose decision-making processes are barely comprehensible. In the worst-case scenario, we could create AI systems that act autonomously - and are not in the interests of humanity.
A cautionary example: the "Paperclip Maximizer" by the philosopher Nick Bostrom.
A seemingly harmless AI is programmed to produce as many paper clips as possible - and ends up turning the entire earth into raw material for it. Sounds absurd? Perhaps. But the scenario shows how important clear Limitations and ethical guidelines are.3
What we can do - responsibility instead of leadership at any price
The race for AI does not necessarily have to end in a loss of control. Companies, institutions and governments have it in their own hands to take countermeasures through clear rules and considered decisions. A key step in this direction: Guidelines for the responsible use of AI develop. Such guidelines help to harmonize technological innovation with ethical principles, safety and social benefits.
However, it is not enough to simply formulate such guidelines, they must also be actively practiced! This is the only way to identify potential risks at an early stage and shape AI development sustainably. A well-founded decision-making process that systematically weighs up both opportunities and potential risks is crucial for this.
Our white paper provides a practical introduction to this topic: Guideline for the development of AI guidelines.
Race along or take a run-up - what really counts in the end?
The AI-The race is in full swing, but it is no longer just about power or markets. It's about how we deal with a technology that will determine our future.
Speed alone is not enough. If you start running without direction and security, you risk losing control. That's why now is the time to stop and ask questions:
Do we simply want to race along or consciously take the next step?
Whether we end up as AI slaves or responsible pioneers will not be decided in China or the USA, but here. With you. With all of us.
It's time to run the race not just faster, but smarter.
