thoughtsparkideas

Artificial intelligence once again dominates the news cycle. But this time, something feels different.

Over the weekend, Anthropic CEO Dario Amodei called for slowing the pace of development at the frontier of AI and putting stronger safeguards around increasingly powerful systems. Days earlier, an Anthropic researcher resigned, publicly warning about the risks of racing toward superintelligence.

Other prominent AI leaders have echoed similar concerns.

Think about that for a moment.

The people building some of the most powerful technology humanity has ever created are warning us about what they are creating.

When the Pioneers Sound the Alarm

Technology companies generally ask governments for fewer restrictions. They want the freedom to innovate, compete and grow.

Today, some of the pioneers of AI are effectively saying something very different: this technology may be advancing faster than our ability to safely manage it.

That is extraordinary.

The warnings are also no longer limited to job displacement, misinformation, privacy or copyright. Some people closest to the technology are openly discussing the possibility that sufficiently advanced AI could eventually surpass human intelligence and become difficult to control.

The debate, therefore, is becoming much larger than technology.

It is becoming a debate about humanity itself.

Who Is Responsible?

I have mixed feelings about these developments.

On one hand, it is encouraging that leaders within the AI industry recognize the risks. Awareness is an important first step.

But something bothers me.

Why are we looking almost entirely toward governments for the answer?

Governments absolutely have a role. Technology this powerful cannot be governed solely through voluntary promises from private companies.

But legislation cannot replace responsibility.

If the people developing AI genuinely believe it could pose catastrophic risks, shouldn’t the industry itself develop a stronger ethos governing how this technology should be developed?

Medicine has the idea of first, do no harm. Engineering has safety standards. Scientific research has ethical frameworks.

What is the equivalent for artificial intelligence?

Regulation without an underlying ethos is simply compliance.

And compliance is a poor substitute for conscience.

An Almost Impossible Problem

I also sympathize with lawmakers.

How do you regulate something changing this quickly?

And there is an even more difficult question.

What happens if one country slows down and another doesn’t?

Artificial intelligence is no longer simply a commercial technology. It is becoming an economic, scientific, military and geopolitical capability.

Move too quickly and we may create something we cannot adequately control.

Move too slowly and someone else may create it first.

Regulate too aggressively and innovation may move elsewhere.

Regulate too lightly and the consequences of failure could affect everyone.

There are no easy answers.

I Keep Coming Back to Consciousness

In several of my earlier articles, I have explored consciousness, thought and humanity’s relationship with technology.

I have wondered whether consciousness is more fundamental than intelligence. I have written about the extraordinary power of human thought—the ability of the mind to travel beyond the physical limitations imposed upon us. And when writing about AI, I have repeatedly returned to one question:

What happens to us as our machines become increasingly intelligent?

The developments of the past few days make me return to that question again.

For thousands of years, humans have created tools that extended our capabilities.

Fire extended our physical power.

Writing extended our memory.

Machines extended our muscles.

Computers extended our ability to calculate.

The internet extended our access to information.

But AI may be fundamentally different.

We are creating a tool that increasingly participates in the very activity that created all those other tools:

Thinking.

And that changes the conversation.

The question is no longer simply:What can we build?

It is:What should we build?

And perhaps more importantly: Do we possess the collective consciousness to decide?

The Intelligence-Consciousness Gap

Perhaps this is the real problem confronting us.

We are rapidly increasing machine intelligence without necessarily increasing human wisdom at the same pace.

Our machines are becoming more capable.

Our institutions are struggling to understand them.

Technology companies are looking toward governments for rules.

Governments are looking toward technology companies for answers.

Countries are watching one another, afraid that whoever slows down first may fall behind. And humanity is somewhere in the middle. This concerns me more than AI itself. Technology has never possessed morality.

Humans give technology direction. Nuclear science can illuminate a city or destroy one. Biology can cure disease or create dangerous pathogens. The internet can democratize knowledge or spread misinformation at extraordinary speed.

Technology provides capability. Consciousness determines what we do with it.

The Next Stage of Human Evolution?

In my earlier writing, I have often wondered whether human evolution is ultimately as much about consciousness as it is about biology.

Perhaps AI will test that idea.

For millions of years, our evolutionary advantage has been intelligence—our ability to imagine, cooperate, solve problems and create tools.

Now we are attempting something remarkable. We are creating intelligence outside ourselves. Perhaps the next test of human evolution is therefore not whether we can create something smarter than ourselves.

We probably can.

The real test may be whether our consciousness can evolve quickly enough to responsibly govern what our intelligence has created.

That requires more than regulation.

It requires some collective understanding of what we are trying to preserve:

Human agency.

Human dignity.

Human purpose.

Human consciousness.

Perhaps even human existence itself.

And that is where I believe we have a problem. We are debating how quickly AI should advance before agreeing on where we want it to take us.

We are trying to write rules before agreeing on principles.

And we are building intelligence faster than we are developing the wisdom required to govern it.

Perhaps the greatest danger of artificial intelligence is not that machines will become conscious.

Perhaps it is that humanity will fail to act consciously while creating them.

The pioneers of AI are asking us to slow down.

Maybe we should.

Not because slowing down solves the problem.

But because it gives humanity something increasingly precious: Time to allow our consciousness to catch up with our intelligence.