AI Can Be a Useful Tool in the Judging Process at the Olympics

There is a fascinating situation developing in the area of figure skating as we move closer to the 2026 Winter Olympics.

Judges will have access to an AI tool to help judge skaters, but this is not meant to take the place of judges, but rather to aid them in their decision-making process.

This distinction seems to mean more than one might initially think.

Problems in Figure Skating Judging

Judging in figure skating has long been a source of controversy.

Rotation, edge quality, landing stability and micro-execution errors all contribute to the complexity of interpreting what a skater has done.

Judges have to interpret all of these aspects of a skater's performance in real-time and judges have differing opinions on the same performance.

The goal of using AI to aid in the judging process is not to allow AI to choose who wins; the goal is to allow AI to accurately measure the technical elements of a performance and provide that information to the human judges.

In other words:

Use AI for measuring the objective aspects of a performance (rotation count, jump height, detection of under-rotation, etc.) and then have human judges evaluate the subjective aspects of the performance (artistic impression, interpretation, emotional connection).

Structural Importance of the Judging Model

Most companies attempting to integrate AI are looking to substitute human workers for AI models.

Replace the junior staff with AI.
Replace the analyst with AI.
Replace the customer service representative with AI.

However, the Olympics' judging model is different.

Use AI to perform precise and repeatable measurements, and have humans evaluate the nuances of each performance.

That is not a philosophical standpoint; it is simply a matter of dividing up the work that needs to be done.

Common Corporate Misconception

Many corporations attempt to assign AI roles that are not structurally feasible or appropriate.

Ask models to:

• Interpret the subtlety of a customer's emotions
• Determine the price of a product based upon the characteristics of the product
• Assess the level of risk associated with an individual transaction
• Make strategic trade-off decisions

When the model does not perform as expected, the corporation is frustrated because it was expecting the model to perform in a manner consistent with the expectations of the corporation.

More likely than not, the problem is not with the model itself, but rather with the confusion of the model's role in the overall process.

AI is useful when the task is:

• Well-defined
• Highly data-driven
• Rules-based
• Quantifiable

However, AI is less reliable when the task involves making judgments based upon values, trade-offs and accountability.

The Olympics provide a clear boundary.
That boundary is the lesson

AI as a Technical Layer, Not a Moral Authority

In figure skating, the AI would not be responsible for determining the artistic impression of a performance.

It would not be responsible for deciding how moving a performance felt.

It would be providing structured technical data to the human judges.

That separation preserves the legitimacy of both the technical layer and the human layer.

Legitimacy is equally important in corporate environments.

If employees believe that AI is "deciding" arbitrarily, they will resist its implementation.

If AI is viewed as a technical support tool that provides data, identifies anomalies, accelerates the analysis process, employees will view it as trustworthy.

The key difference is architectural clarity.

Implications for Corporations

The correct question is not:

"How much can we automate?"

It is:

"Where should automation stop?"

Many corporations deploy AI without clearly defining the lines of responsibility.

Who is accountable if the AI model produces incorrect results?
Who has the ability to overrule the AI?
Which decisions should remain explicitly human?

Without clearly defining those parameters, the integration of AI appears to be chaotic.

The Olympics' judging model provides implicit answers to those questions:

AI measures.
Humans decide.

Deeper Insight

Full substitution of human workers with AI typically represents the largest barrier to AI adoption in corporations.

Corporations frequently pursue cost savings prior to establishing a clear understanding of how AI will fit within their overall capabilities.

The Olympics suggest an alternative approach:

Let AI improve the precision of technical measurements.
Allow humans to utilize that increased precision to improve their decision-making processes.

That approach maintains both the benefits of speed and accountability.

Practical Takeaway

AI performs best when it:

• Reduces the measurement error
• Increases the processing speed
• Identifies hidden patterns
• Provides standardized analysis for repeated tasks

AI performs poorly when it is given unclear authority.

The Olympics illustrate a balanced architecture:

• Clear division of labor
• Explicit role boundaries
• AI as a technical infrastructure
• Humans as decision makers based upon context

That is not sexy.
But it is sustainable.

by Eugene Baiste, AI Capability Architect at Skill Sonic