Sylvain Charlebois: AI can make our food system more efficient, but if we’re not careful, also more expensive

The debate over artificial intelligence has moved quickly from fascination to anxiety. We are asking what AI can do, who will control it and whether governments can keep pace. Bill Gates’s warning that in the wrong hands, AI could cause a billion deaths captures the scale of that concern. It is a warning about catastrophic misuse, rather than a prediction that such deaths are inevitable.

When it comes to the food supply, the questions are more immediate. Will AI help make groceries more affordable, or simply make companies better at determining how much each of us is willing to pay?

McDonald’s recently announced US$8.5 billion (C$12 billion) in support through 2036 for restaurant modernization, technology deployment and operational improvements, including AI. The entire amount is not an AI budget, but the direction is clear: the industry sees productivity gains in technology.

Compare that with Wendy’s experience in 2024. Its discussion of dynamic pricing triggered a backlash over the prospect of paying more for lunch during busy periods. The company subsequently clarified that it would not introduce surge pricing. Consumers heard a threat to affordability, and the response was swift.

McDonald’s has also faced scrutiny over its pricing tools. The company says its system provides recommendations, but franchisees make the final decisions and prices are not set according to an individual customer’s willingness to pay. The debate shows how sensitive the relationship between technology and food prices has become.

The industry should take that lesson seriously. Public acceptance depends on what AI is used for, how clearly companies explain it and who benefits.

The scale of investment is already remarkable. In September, Saskatchewan announced an expansion of Bell’s AI data centre project, bringing the estimated capital investment to more than $50 billion and establishing its headquarters in the Regina area. AI is becoming infrastructure, with consequences for energy, investment and economic development.

Interest is also growing across sectors connected to Canada’s food supply chain. Statistics Canada figures released Oct. 1 show that 13.2 per cent of businesses in agriculture, forestry, fishing and hunting plan to use AI over the next 12 months, up from 4.9 per cent a year earlier. In accommodation and food services, the share increased from 14.8 to 17.4 per cent.

These figures measure intentions, rather than actual adoption. The categories also extend beyond food. Still, they point to growing interest in industries facing persistent pressure to improve productivity, manage costs and find workers.

Technology has been changing food production for years. Precision agriculture can help farmers target irrigation, fertilizer and crop protection more accurately. AI can strengthen those decisions by identifying patterns and anticipating needs. In distribution, forecasting and automation can improve inventory management, delivery schedules and the movement of perishables.

Much of this work happens beyond the consumer’s view. A more efficient warehouse rarely attracts the attention that a price change does. That is where the debate becomes more difficult.

Location pricing, dynamic pricing and surveillance pricing are different practices. Restaurants and stores have long charged different prices at their various locations due to differences in rents, labour costs and competition. Dynamic pricing changes prices over time. Surveillance pricing uses personal information to tailor the price offered to a particular shopper.

Finding a different price for the same meal at two restaurants does not, by itself, prove that either restaurant is using AI or profiling customers.

But using someone’s browsing history, purchasing habits or inferred income to charge more for food raises a fundamentally different concern. Imagine two people buying the same basket of goods at the same moment, with one paying more because an algorithm believes that person can afford it — or has fewer alternatives.

That should trouble anyone concerned about food security.

Maryland has already enacted legislation restricting surveillance pricing in groceries. Canadian policymakers should establish clear prohibitions against using personal data to impose higher food prices.

Those protections should preserve useful practices, including transparent promotions and markdowns that help sell food before it spoils. An electronic shelf label can reduce administrative work and make discounts easier to offer. The concern lies in how a pricing system treats people.

We tolerate fluctuating prices for flights and hotel rooms, but food carries a different social obligation. Households can postpone a holiday. They cannot postpone eating indefinitely.

AI could help make our food system more productive. Better forecasts can reduce waste. More efficient operations can ease cost pressures. Yet efficiency does not automatically translate into lower prices. Competition will help determine how much of the benefit reaches consumers.

Consumers, meanwhile, will have tools of their own. AI could make it easier to compare prices, identify worthwhile promotions, plan meals and avoid purchases that offer poor value. A retailer may become better at predicting what a shopper will buy, while that shopper becomes better at finding another place to buy it.

That contest will create opportunities and expose weaknesses. Businesses that ignore AI risk losing ground. Businesses that use it carelessly risk losing trust.

Food companies should pursue innovation with a clear understanding that consumer confidence is an economic asset. Once damaged, it is expensive to rebuild.

AI should help the food industry produce more, waste less and serve people better. Using private information to determine how much extra a household can be made to pay for dinner crosses a line.

Our food system needs greater intelligence. Consumers deserve fair treatment at the checkout.

National Post

Sylvain Charlebois is director of the Agri-Food Analytics Lab at Dalhousie University, co-host of the Food Professor Podcast and visiting scholar at McGill University.