Do we need AI regulation?

AI has the potential to become a killing machine and needs to be regulated. That is one rather hyperbolic argument put forward by some when it comes to AI regulation. But just how true is it?

If you were to believe experts such as Mustafa Suleyman and former Google CEO Eric Schmidt, the potential likelihood of AI developing in such a manner is quite high without proper regulation. To counteract this, they propose the formation of an international panel on AI safety to regulate the technology. This panel would be made up of experts that would be empowered to objectively inform governments about the current state of AI capabilities and make evidence-based predictions about what could come next.

Whilst it might seem self-evident that politicians need to know about what it is they are regulating and the challenges that this will cause for them, some healthy scepticism is needed. After all, the alarm bells about AI are not being raised by the public or by politicians but by those in the AI business themselves. 

Subsequently, one must wonder if those calling for regulation within the industry are saying more about themselves and their colleagues than about any actual serious threat. After all, governments or hostile actors will only be able to use AI for nefarious purposes if the developers enable AI to be able to carry out such processes. 

At present, beyond being used as a language model in things such as ChatGPT or in medical experiments or in some capacity in self-driving cars, AI’s abilities have been found to be wanting.

To make matters even more interesting, there exists no concrete definition of what AI is. Without a clear definition of what AI is and what it isn’t, how can the experts expect governments to effectively regulate its development? In this instance it does seem as though they are putting the cart before the horse.

This has not stopped governments from introducing legislation to regulate AI, or even starting discussions with their allies.  Which raises concerns about just how much understanding exists amongst the governing elite about what they are currently doing.

What should perhaps be done is to reach an agreement about what AI is and where its limitations currently lie. From there, a proper strategy can be developed on a national and international basis for regulation. Anything else would be half-cocked.