Generative AI flew through the hype cycle at breakneck speeds the past year to arrive in 2024 as ‘the year of reckoning’, according to Elizabeth Lopatto at The Verge. There’s been a lot of discussion around what generative AI could do, but this year leaders should focus on “what AI can actually do right now.” In order to achieve that, leaders in the visitor attraction sector need to understand its possibilities, limitations and practical applications in order to govern how it’s infiltrating the attraction.   

Artificial Intelligence (AI) disrupts at the heart of a leadership team’s accountability on two strategic fronts: opportunity and risk. For generative AI, this is about using the technology to create ‘new’ content in the form of text, images or even music, rather than the traditional realms of AI which excel in logical or analytical tasks. It’s a tectonic shift potentially as significant as digital to the visitor attraction sector – certainly to back of house operations and even for visitors and their experiences. What makes this challenging is that AI is so broad in its technologies, applications and ramifications. It’s also a change that’s already here. In many visitor attractions organizations, AI will already be a part of how individuals and teams operate, potentially unknown to its leadership team.

Even though the age of the AI generated experience in a visitor attraction is still likely some years away, experience teams may already be using generative AI to generate everything from catchy names for their next event through to background images used in an exhibition. In the shorter term, visitor attractions will find other opportunities of where and when, and how much to invest, as the industry navigates the discoveries and realities of generative AI in cultural and commercial attractions.  

Back of house functions such as customer services, finance, marketing and human resources are prime targets to become AI empowered, and potentially using generative AI more specifically. Examples such as forecasting using machine learning, evaluating customer feedback using natural language or automating customer messaging through generative AI are all opportunities where routine tasks can be done more efficiently and potentially better than a human hand. For some, these tasks could otherwise never be done affordably or objectively at scale, such as transcribing the sentiment of customer calls or processing collection images for quality control.

AI is also transforming the workforce. Roles such as human resources might use generative AI to draft a job description, or a marketer to create ad copy. This presents two different fronts for companies to manage: enough education and encouragement for individuals to get the most from AI to make them more efficient and effective, but enough guardrails, such as enterprise subscriptions to ringfence data, to protect the attraction from the risks of this use. Staff need to be knowledgeable around topics such as data sensitivity, intellectual property and copyright, plus the tasks for which certain types of AI is inappropriate for, or where it can introduce bias. In uninformed or inexperienced hands, generative AI in particular can create a false confidence in the technology’s ability, then used without proper checks on its accuracy. For example, it can be terrible at basic math!  

At Dexibit, we’ve long used AI for several purposes, such as predicting visitor behavior using machine learning for forecasting metrics like visitation and interpreting insights using natural language for sentiment on visitor feedback. Recently with generative AI, we introduced our new AI Query Assist, a magic wand which turns a user’s words in plain English into a query using Dexibit’s own query language, DxQL. It’s part of our vision to democratize data from the front door to the board room, no matter a user’s digital confidence or data literacy – now everyone can be a data analyst! It’s a small but mighty and highly practical example of how generative AI can be used to power tasks within a job. However, it also comes with the usual fineprint of generative AI: the output is usually only as good as the prompt and it’s a ‘probabilistic product’, meaning it will usually get the right result but relies upon the user to validate that’s the case.

When it comes to prudent risk management in the age of AI, it is important for visitor attractions to manage additional risks from unintended consequences such as the potential for discrimination, abuse and fraud. For example, an increasingly common issue in the attraction sector is the use of generative AI by bad actors in review ransom scams, where AI is used to create a high volume of negative but realistic looking reviews targeting a particular location on social media. By the time the review channel steps in, the revenue damage can already be significant. 

Visitor attraction leadership teams also face a changing regulatory landscape, especially for federal organizations or those operating internationally. Governments the world over are attempting to legislate or at least introduce voluntary codes of conduct for ethical and societal implications including complex demands for transparency and explainability in a world where most AI operates as a ‘black box’. For example in the US, this is the Blueprint for the AI Bill of Rights or in the EU, The AI Act. 

In order to exploit these opportunities and mitigate the inevitable risks of AI, executives in visitor attractions must stay apprised of these rapidly evolving technologies, their applications and the market’s reception. Whilst a diverse team with specific expertise is a start, the broad ramifications of AI demand a level of awareness in digital literacy from all leaders and the cascade of this knowledge through the organization. Together, the visitor attraction will need to quickly and continuously extend risk management policies, processes, procedures, practices, accountability and training for a world ever changed by AI.