AI Makes More Possible. It Doesn’t Make Everything Advisable

The recent warnings from Dario Amodei, Sam Altman, and other AI leaders are noteworthy. Not because every scenario they describe will materialize, but because leading developers are now publicly asking whether we have reached a point where making AI more capable faster does not automatically mean making progress.

This is the essential point of discussion.

In many companies, speed has become a success metric in its own right. The faster you develop, scale, and bring something to market, the more innovative you are perceived to be. With AI, that logic only goes so far. As systems become more autonomous, measuring their capabilities alone is no longer enough.

The critical question is:

What happens when a system becomes capable of doing something we did not anticipate?

In digital growth and transformation initiatives, I have seen that organizations rarely underestimate risks because they lack technical understanding. More often, they underestimate them because accountability is not clearly defined. As long as everything works, this hardly gets noticed. Once a system starts acting on its own, however, a technical issue can quickly become a leadership issue.

That is why I consider independent security audits and clear boundaries necessary—not as a brake on innovation, but as an integral part of professional innovation.

We should also be careful with “grand” doomsday scenarios. They generate attention, but do not necessarily lead to better decisions. The more relevant risks are much closer to home: for instance, an autonomous agent exploiting a security vulnerability, accessing a system it was never supposed to reach, or making a decision for which no one is clearly accountable.

That may be less spectacular than the idea of AI taking over the internet. For a company, it can nevertheless become a real problem.

I therefore do not believe we need to slow down AI across the board. We do, however, need to be more deliberate about where speed makes sense—and where oversight, testing, and human accountability need to set the boundaries.

Progress does not mean implementing every technical possibility as quickly as possible. Progress also means recognizing which possibilities we are not yet ready to use responsibly.

This is how sound leadership in AI initiatives will be measured.

Disclaimer: All information provided in this article is for general informational and positioning purposes only. It does not constitute legal, regulatory, tax, or financial advice, nor does it serve as an instruction or solicitation for any transaction. Content is provided “as-is” without warranties of accuracy or completeness. Any views or opinions expressed belong solely to the author and do not necessarily reflect the official policy or position of the operator of this website, the author’s past or current employers, or any affiliated organizations. Readers are solely responsible for conducting their own independent due diligence before taking any action.