Artificial intelligence is gradually transforming event management: content generation, automation, personalized experiences, and support for organizers. But these digital applications also have an environmental footprint. Electricity consumption, water use, data centers, model selection, and query complexity—the impact of AI depends on many factors. So, how can we reap the benefits of AI in event management while adopting a more responsible approach?
Artificial intelligence is no longer just a future-oriented topic for event professionals. It is beginning to play a tangible role in various stages of event planning.
In particular, it can help generate or reword content, organize certain information, automate repetitive tasks, personalize user journeys, or assist organizers in setting up their events.
At AppCraft, this approach is reflected, for example, in the use of AI to assist teams with certain configuration tasks. The goal is not to replace the event organizer, but to allow them to spend less time on repetitive tasks and more time on designing the participant experience.
This development, however, raises a new question: What is the environmental impact of these uses of artificial intelligence?
Interacting with artificial intelligence seems completely intangible. Yet behind every query lies physical infrastructure.
When a user queries an AI model, the request is processed by servers equipped with specialized processors, housed in data centers. These facilities require electricity to perform calculations and to power and cool the equipment.
The environmental footprint of AI use can thus encompass several aspects: energy consumption, CO₂ emissions, water consumption, and the production of IT infrastructure.
We must therefore avoid two oversimplifications: viewing AI as “intangible” and having no impact, and assigning a single figure to all uses of AI.
Talking about “the impact of an AI query” as if it were a universal value is misleading.
In his article “The Environmental Footprint of AI (Part 1),” mAIjin demonstrates that the orders of magnitude can vary considerably depending on the use case. The article specifically cites several key factors: the model used, the length of the context, the use of reasoning, agent-based applications, the region where the system is hosted, and the energy efficiency of the data center.
A simple text request and a complex video generation task therefore do not place the same demands on the infrastructure.
This is a key point for seriously addressing the issue of responsible AI: before comparing figures, we must understand what they measure and under what conditions they were obtained.
Check out: The Environmental Footprint of AI (Part 1) by mAIjin, which details the various factors that influence the environmental footprint of artificial intelligence use.
The nature of the task at hand plays a major role.
According to the orders of magnitude compiled in mAIjin’s article, a short text prompt is a far cry from a full-fledged agentic task or video generation. For example, the document presents a range of 0.2 to 0.5 Wh for a short text prompt, compared to 10 to 100 Wh for a full-fledged agentic task and 9 to 400 Wh for a 5- to 8-second generated video.
These values should not be interpreted as universal constants. Above all, they illustrate a reality: referring to “AI” as a homogeneous concept masks considerable differences among applications.
In the events industry as well, it is therefore important to consider not only whether to use AI, but also how it is used and the value that each application actually creates.
The type of content generated is not the only factor.
The amount of information fed into the model, the complexity of the reasoning required, or the number of actions needed to complete a task can significantly affect the resources used.
A short request requiring a simple response is not comparable to the analysis of a very large document or to an agent tasked with automatically performing a series of operations.
For businesses, this distinction becomes increasingly important as AI is integrated into business processes. The question is therefore no longer just “Are we using AI?”, but also “For what tasks, with what level of complexity, and to what end?”
The goal of a responsible approach is not necessarily to abandon artificial intelligence.
Rather, the goal is to avoid using complex models for tasks that don’t require them and to focus AI on applications that can generate real operational benefits.
In the events industry, this can mean using AI to simplify a complex setup, speed up event preparation, automate certain repetitive tasks, or make it easier to customize a participant’s experience.
This approach makes it possible to move from using AI “simply because it’s available” to using it in a way that is guided by usefulness, effectiveness, and relevance.
The environmental impact of a digital tool should not be analyzed entirely in isolation from the process it transforms.
For example, a centralized event platform can minimize certain manual tasks, reduce scattered communication across multiple tools, and facilitate the digitization of certain processes.
The use of digital forms, targeted communications, mobile-friendly programs, QR codes, and digitized documents can help reduce the need for certain physical materials and unnecessary tasks.
This does not mean that digital technology has no impact. It means that a responsible approach must examine the entire system and the trade-offs made, rather than systematically pitting digital technology against the environment.
AI is just one component of an event’s environmental footprint.
Transportation for participants, lodging, food and beverage services, equipment, energy, services, and digital usage all contribute, to varying degrees, to its overall environmental impact.
That’s why AppCraft also lets you calculate and track an event’s carbon footprint, to help organizers better identify the main sources of emissions.
This approach makes it possible to put each use case into context and avoid focusing all efforts on a single, visible aspect that may be secondary in relation to the event’s overall footprint.
It is justified when it provides a real operational benefit that is proportional to the complexity involved. Three questions help determine this: for what task, with what level of complexity, and for what value? Simplifying a complex configuration, speeding up setup, automating repetitive tasks, or customizing a participant’s experience are relevant use cases. The order-of-magnitude figures cited by mAIjin demonstrate the value of avoiding complex models or agents for simple needs: 0.2 to 0.5 Wh for a short text prompt, compared to 10 to 100 Wh for a full agent-based task. These figures are indicative and vary depending on the model and infrastructure.
By treating it as just one component among many, rather than in isolation. Participant transportation, lodging, food service, equipment, energy, services, and digital usage all contribute, to varying degrees, to the overall carbon footprint. Measuring the entire picture helps identify the main sources of emissions. This prevents us from focusing our efforts on a highly visible but potentially secondary source. A tool for calculating and tracking carbon impact, such as the one offered by AppCraft, helps put the use of AI into context.
Not automatically: Digital technology also has an environmental footprint (electricity consumption, water, data centers, infrastructure manufacturing). However, it can reduce other forms of waste. A centralized platform minimizes certain manual tasks and the scattered exchange of information across multiple tools. Digital forms, mobile apps, QR codes, and digitized documents can replace physical media. The right approach is to examine the entire system and the trade-offs involved, rather than systematically pitting digital technology against the environment or viewing AI as intangible and impactless.
