We thought we knew where the internet would take us

In the late 90’s I spoke at conferences across the world about the Internet and how it would change all our lives. I was running a company that was building the internet out across Europe and Asia. I worked with experts in the field, and I was often described as one myself.

We were certain that we knew the direction. We thought we understood the consequences. We were filled with enthusiasm and passion. Build it and they will do great things with it, we said. We imagined worldwide communication spreading knowledge, truth flourishing because it could no longer be hidden. We saw the green shoots. The vision was becoming real and in many ways, it still is.

But then the reality arrived. Alongside the great connections came the endless pictures of cats, the pornography, the scams and the harm. Yes, there was more communication, but not all of it was good. Truth was twisted. Bad actors flourished. Attempts at control and regulation failed, and continue to fail. The internet has changed the world, but how much – and in what ways – has it truly changed it for the better?

the parallels with ai are uncomfortable

The parallels with today are obvious. We are told with certainty how AI will change everything. There are already examples of great good and examples of great harm. What we cannot say with confidence is where AI will actually take us.

And that uncertainty raises the uncomfortable question: what should we be doing differently this time?

what could we have done differently with the internet?

Looking back, there are questions that challenge all of us who were part of that first wave:

• Did we ask enough difficult questions? We assumed that if the infrastructure was built, it would inevitably be used for good. What if we had interrogated those assumptions more harshly?

• Could we have built in safeguards earlier? Regulation is now playing catch up – slowly, clumsily – and yet the damage is already done. Could we have demanded smarter oversight from day one?

• Were we naïve about human behaviour? We underestimated the capacity for distortion, harm and manipulation. The same idealism colours the AI conversation today.

what lessons should we take into the AI age?

We do not know where AI will go, anymore than we knew where the Internet would go. But we do know that the stakes are just as high, perhaps higher. The lessons we might consider now:

• Be sceptical of certainty. Anyone who says they know exactly how this will play out should be challenged.

• Design for misuse, not just use. Assume that bad actors will find ways to exploit the technology and plan for that.

• Shape the rules before it’s too late. Waiting for harm before regulating leaves us constantly reacting, never leading.

AI, like the internet, will be a mirror. It will reflect our brilliance and our flaws. The question for leaders, innovators and policymakers is whether we will repeat the same mistakes – or finally learn from them?