Organising events is complex work that demands planning, coordination and attention to detail. With the arrival of artificial intelligence the scope for optimising those processes has grown exponentially: to name just a few cases, there is the automation of repetitive tasks, which lets organisers save time and resources and concentrate on higher-value work, or predictive analytics and the rapid processing of large volumes of data, which allow the event experience to be personalised to levels never seen before, raising engagement and making high levels of satisfaction attainable.

Adopting AI in events, however, also brings significant ethical and legal questions with it — ones that do not yet seem to have been brought into focus as clearly as they should. The first concerns transparency: what is being done with AI and how it is being done, and understanding the mechanisms by which this extremely powerful tool works. Without wanting to venture into very broad territory, we would point you to the EU Artificial Intelligence Act, the European Union document we would recommend reading in any case. More simply, what we want to do here is raise some questions we think useful for approaching the use of this new tool in organising and running events with a little more caution — questions revolving around the collection and processing of data, respect for privacy and compliance with the regulations governing both.

One issue that matters most to us, because we can see it affecting an ever larger share of the people organising events on our platform, concerns the processing of data — particularly the personal data of event registrants — by software that uses artificial intelligence. Certainly, in these cases processing large volumes of data becomes more agile, faster and more precise, and lets the extracted information be put to work immediately, to the benefit of data-driven marketing and of future planning. And all of that is good, very good, particularly in a world moving ever faster and therefore needing answers that are ever more immediate but still correct, so that resources and opportunities are not wasted.
What we have happened to observe, though, is that these tools are sometimes relied on before any adequate analysis of the risks their use carries, and before the necessary precautions have been taken.

Let us give some examples.

Who owns the data once it has been handed over to an artificial intelligence system for processing? Where will it be stored and how will it be used? We know it will very probably be used to train the AI further, to help it grow and improve its performance. But what is the cost of that? And the risks? And the responsibilities of whoever physically handed the data over, especially when it belongs to third parties? How were those third parties informed, and how is their consent managed? All questions we are only now beginning to ask ourselves — and that we must begin to ask, precisely because the subject moves fast and the technology faster still.

Think of a simple and fairly common case. People increasingly attend events to widen their personal and professional networks, which means event organisers are interested in equipping themselves with systems capable of matchmaking and of making networking between people easier. Artificial intelligence can be a fantastic ally in that work, because it can process large volumes of information quickly and precisely and return results that are immediately usable, genuinely helping Paolo meet Carla and Federica get to know Raffaele. But where is the data belonging to Paolo, Carla, Federica and Raffaele sent in order to be processed by the AI? Where will it be stored? Will it be accessible, retrievable and amendable as the regulations require?

Asking these questions before — rather than after — deciding to use these resources can make all the difference in the face of any legal action, and can stop event organisers finding themselves, possibly alone, having to provide answers when the real parties to ask are somebody else: the very service providers who are hard to reach when they are not actually hard to identify.

Certainly, the main institutions — starting with the data protection authorities — are gearing up to issue guidance, set limits and provide for penalties. All of it right, timely and useful, because it helps establish guidelines to follow in trying to govern the complexity. What has to accompany it, in our view, is a shift in mindset further upstream: one that lets us see risk before it becomes a threat, and that spares us both from shutting the stable door after the horse has bolted and from denying ourselves opportunities out of fear. The other day someone was talking about an “AI licence” to be earned before you are allowed to use these tools — and who knows, they may have had a point.