Mr. Denis Hannigan:

I acknowledge and reiterate the point that it is a very nascent technology. It is important we are talking about and recognising many of these risks and issues. We are seeing that these require new ways of working and new approaches to mitigate and manage those effectively. To use one client example, if I may - I know it is not quite the scenario alluded to in the question - we are working with an organisation on its chatbot capability and infusing GenAI into that to make it much more engaging for the customers who use that chatbot. That has recently gone live over the course of the past six weeks. It has required a very different way of working for that organisation. This technology is non-deterministic, which means that if you ask it the same question multiple times, you will get different answers on some occasions. You need to rethink how you will test and how you will make sure you are comfortable with the guardrails around that. Our teams had to do a lot of fine-tuning of that particular model, putting in 150 FAQs and putting the right tone and controls in place to ensure this chatbot would respond in a manner that was in keeping with what the organisation was hoping. One of the learnings was that the hardest aspect of the entire project or initiative was bringing some of the stakeholders from privacy, security and risk on board on that journey and ensuring all different stakeholders were comfortable.

That organisation now feels better equipped as it looks to scale up into other areas and the feedback has been positive so far. The key point is that it is a new way of working. It requires a new mindset and new approaches and is something we need to take seriously.