Richard Bruton (Dublin Bay North, Fine Gael) | Oireachtas source

There is a major controversy about ChatGPT, the artificial intelligence platform. The case is made that there is no oversight or transparency regarding the software underpinning it, the algorithms it uses and the principles it applies. I saw an article contrasting it with airline autopilot systems, which are regulated in immense detail and with every little bit exposed to public scrutiny. With regard to risk response and the mitigation of risk, how will this legislation apply to an operation like the one in question? We already hear allegations that it could be used in criminal activity by replicating people’s voices, speech and so on. This is the scale of the threat that people are beginning to identify on the horizon. Is the legislation fit for purpose in taking on the emerging threats, or will we be back in a year saying we need to think again because the platforms have got more powerful?

Embedded in the legislation that went to the media commission was the idea that there is a very strong obligation on an entity to make sure its codes and the principles it applies are soundly based. To fail to do that is in itself an offence. A commissioner can go in and say the entity does not have adequate procedures. I wish to understand this because it is a tricky area. We do not fully understand artificial intelligence as legislators, no more than the delegates do as regulators. Will we be back in a year saying we do not have the framework for ChatGPT or the next incarnation of artificial intelligence?