Part 1: (Artificially) Intelligent Route Setting.
How artificial intelligence and other technologies are playing a role in the development of climbing gyms.
There is very little data on AI adoption in the climbing world, and so I’m asking my readers to take this very short survey. Click the link above to participate. Results will be out in a few weeks!
Introduction
I’m sure by now you’ve heard of Artificial Intelligence. AI has been the buzzword of 2023/2024, and industries are still scrambling to understand how it will affect them. Triggered by the launch of ChatGPT, we’ve been bombarded with new ways AI can solve complex problems, boost sales, and make our lives more efficient.
In my conversations with industry professionals—be it instructors, technical advisors, outdoor guides, or business owners—the topic of AI is surfacing more frequently.
What excites me most is how these different stakeholders see these technologies in completely different ways: outdoor guides query what will happen when people can climb Everest in their living rooms using a VR headset; business owners want to leverage AI to better understand customers and increase footfall in gyms; technical advisors think about ways that image recognition can create warning systems to reduce the number of accidents.
In this article, I’ll look at how AI may disrupt route setting, and what this means both for gyms and the route setting profession.
Image generated by ChatGPT. Oh, the irony.
What is AI?
AI is defined as “the theory and development of computer systems able to perform tasks normally requiring human intelligence, such as visual perception, speech recognition, decision-making, and translation between languages.”
Depending on whom you speak to, this definition might conjure up images of a Hollywood-style humanoid machine or, more commonly nowadays, a ChatGPT-style chatbot. However, even if you have never used a chatbot, it is likely that you encounter AI almost daily - think music recommendations on Spotify, or targeted ads on Instagram.
AI systems are essentially complex models that utilise machine learning (a fancy term for statistics applied to lots of data) to make predictions and perform tasks. Ever wondered why Youtube just keeps suggesting another Magnus Midtbø or Wide Boyz video?
In this context, I’ll mostly be discussing Generative AI and its ability to create new ideas and content based on user prompts.
Machine Route Setters
Route setting is perhaps the most common discussion point when it comes to the use of AI in climbing.
You may have seen the Catalyst Climbing and/or Wide Boyz videos in which they test the capabilities of AI to develop new boulder problems.¹ In short, the videos make the AI output seem basic, and frankly, quite useless. The human route setter is still required to provide significant input to interpret the instructions given by the chatbot, and the boulders often end requiring significant modifications to make them even possible to climb (perhaps the chatbots are setting for humanoid robot climbers).
However, these videos misrepresent AI’s possibilities by not fully considering how Generative AI works—its outputs depend on the data it’s been trained on. If I ask ChatGPT to create an image of a car, it can only do so because it has learned the features of a car (body shape, four wheels, steering wheel, etc.).
Similarly, we can only expect AI to generate high-quality climbing routes at a particular grade if it has been trained on a variety of routes at that grade and understands the necessary components.
Fortunately, apps such as Kilter Board and Griptonite have been tracking climbs for years, so the data exists. The bottleneck is more likely to be convincing the owners of this data to allow it to be used for training AI route setting models.
Feedback Loop
As technology advances, AI won’t just create new routes, it will also generate feedback systems that allow setters and gym managers to immediately evaluate the quality of routes.
No longer will gyms need to install feedback boxes for customers to give their opinion on routes; image recognition technology will indicate which routes are most popular, where climbers fall off most frequently, and which get completed too frequently for their difficulty grade. If you think this all sounds too futuristic, check out Techtopia’s Higher Eye product, announced in 2021.²
Understanding how route setting influences customer behaviour will become crucial for designing spaces that best serve climbers.
Economic Implications
Competition is fierce when it comes to hiring the best setters. As in any industry, financially successful competitors can pay the most to attract the best talent, creating a virtuous cycle.
On the one hand, increased use of AI in route setting has the potential to increase the supply of top quality route setters in a short period of time.³ This could prove massively beneficial to struggling gyms, particularly those in areas with a shortage of setters, allowing them to level the playing field against their more successful rivals.
At the same time, with AI acting as an efficiency-boosting-assistant, a single setter may be able to put up the same number of lines in a day that previously would have required multiple setters. Both these forces will have a downward impact on route setter wages.
On the other hand, greater awareness of the impact that high-quality setting has on a gym’s bottom line might increase demand for these roles, albeit with a shift towards analysing and refining routes based on AI-generated feedback rather than human route setters creating new lines.
Additionally, such technological advancements could open up new markets for route setters in regions where the profession is currently underdeveloped by lowering the barriers to entry.
Conclusion
AI still has some way to go before it becomes a staple tool in route setting. AI models will need to be trained on masses of previously created routes in order to generate new ones effectively.
Once this hurdle is crossed, however, the economic implications for gyms seem fairly clear. Assuming all gyms have access to this technology (a sensible investment, in my view), competition in the quality of route setting will level out, leaving gyms to compete on other elements of the customer experience.
As for route setting itself, AI is unlikely to kill the profession. At a minimum, humans will still be required to physically install the routes. Instead, the nature of the route setter’s role will become more analytical, evaluating how climbers interact with routes and adjusting them in a way that most benefits the gym.
1
Wide Boyz: https://youtu.be/rGRK5aJE5xw
Catalyst Climbing: https://youtu.be/flaykeZyACw
2 https://techtopia.eu/higher-eye/
3 Relevant qualifications will still be needed from a health and safety perspective.