Post by Ifeanyi Uddin (@IfeanyiUddin)

We all agree that we need to design guardrails for how artificial Intelligence (AI) will interact with humanity.

Yet, the pace at which regulations are finally agreed and implemented, in democracies, will not do for this purpose. New AI models with mind-boggling possibilities are released even before the promises and threats from older models have been fully unpacked.

If the regulation of this new technology is not to continue to pay catch-up, then, clearly the need is for a toot-and-branch revamp of how we agree regulations. The temptation is to shorten time-to-market, in order to whack more moles faster.

This would be far easier for autocracies to do. But also far more dangerous. Today, by far the biggest challenge from the new technology is not how few the hands that control it are. The new technocracy threatens the established order in this narrow sense.

But there is a bigger problem. The danger that resources will be re-allocated from society’s retail end – via the job losses that AI models threaten – up market.

To the new technology and business elite that will arise on the back of the AI revolution, any solution to the question of the technology’s regulation that adds an unrepresentative regulatory elite will only compound matters.

How do we solve the dilemma presented by the need for a broad-based regulatory process that can properly husband AI’s pell-mell rush towards recursive self-improvement?

Any LLM willing to give this task a shot?

0 likes · 0 comments · 0 shares