The AI in the line

Many Canadians have set their clocks so they can get online in time to snag concert tickets, campsite reservations or swimming lessons for their children. Miss the window, and you might have to wait until next season — or next year.

But what if you could use an AI bot to wait in line on your behalf?

The Globe and Mail recently ran this experiment and found that it was remarkably easy to snag a high-demand booking as soon as it became available.

We asked Keiwan Wind, an assistant professor of Information Systems, to talk about the potential impacts of autonomous AI on market access and how to redesign unfair systems.

How widely is AI used to make digital reservations?

We do not yet have good evidence to say how widespread AI agents are across reservations such as campsites, swimming lessons or summer camps. The Globe and Mail experiment is important not because it proves widespread use, but because it demonstrates how dramatically the barrier to entry has fallen.

What previously required considerable programming expertise can increasingly be created by describing a task to an AI. As agentic AI develops, these systems will not simply click faster; they can potentially search continuously, make decisions, adapt their behaviour and transact on our behalf.

What does this use of AI mean for equitable market access?

This is where I see the greatest concern. First-come-first-served systems were never completely equitable. They already favour people with flexible schedules, digital literacy, fast internet and the ability to sit in front of a screen at exactly the right moment. AI can amplify those inequalities enormously.

It can also make locality increasingly irrelevant. Imagine an affordable campground, recreational program or other scarce service whose prices partly reflect the incomes and conditions of the surrounding community. AI agents can continuously discover such opportunities for people anywhere in the world. Wealthier outsiders, equipped with better technology and greater purchasing power, could increasingly compete with local residents for resources in their own communities.

The result could extend beyond difficulty finding a campsite. Increased external demand can contribute to higher prices, secondary markets and speculative behaviour, while putting additional pressure on local infrastructure and the environment. We have already seen similar dynamics in housing and short-term rentals: digital platforms can connect local, physically constrained resources to global demand.

There is also a potential reinforcing loop: as some people use AI agents and gain an advantage, others feel compelled to use them simply to remain competitive. Eventually, what began as an advantage can become almost a requirement for participation.

How does bot detection help to distinguish between the legitimate use of AI bots in virtual queues and bad actors?

Bot detection is useful, but I would be cautious about treating it as the solution. Increasingly sophisticated agents may learn to behave more like humans: varying timing, using different identities or accounts and distributing transactions in ways that make coordinated activity appear to be many independent individuals. This creates another potential arms race: platforms improve detection, agents improve deception, platforms improve detection again and so forth.

More importantly, detecting whether something is a bot does not answer the ethical question.

The danger is also much less visible (if not invisible) than in the physical world. If one person stood in a queue holding 200 spots for a reseller, it would immediately attract public attention and raise questions. But when the same behaviour happens through accounts, servers, wires and waves, it can remain largely invisible.

What steps can consumers take to ensure they have fair access to goods and services? How are regulators protecting Canadians’ market access?

Consumers cannot solve a system-level problem individually. Asking everyone to acquire a better AI agent simply creates a technological arms race.

The more important response is redesigning how scarce resources are allocated. Rather than relying on first-come-first-served, weighted systems could consider factors such as previous unsuccessful attempts, accessibility needs and, where appropriate, distance from the resource (i.e. a valid residential address).

This does not mean excluding visitors. A portion of access could be protected for local communities while keeping the resource broadly available. This is particularly important when local people bear the congestion, environmental pressure and other costs created by increased outside demand. The goal should be to balance openness with local access, rather than allowing technological advantage and purchasing power to determine who gets access.