The legal considerations of AI-powered food outlets

News of AI-powered drive-thru coffee shops coming to the UK made headlines last month, with plans afoot for an operator to open 22 outlets over the next year.

Shoppers in London have already been among the first to shop at unstaffed, digital-first supermarkets, which received their fair share of both applause and criticism, however the idea of an AI-powered food and beverage (F&B) vendor, which will be staffed mainly by robots handling and serving hot drinks and food, brings with it a whole host of new legal challenges.

While such new concepts are exciting, the introduction of AI doesn’t necessarily eliminate risk, which in itself presents a whole host of new legal considerations for operators to consider.

Contracts: Commercial v Employment

If you replace your people with robots, you are effectively replacing HR issues with commercial contract disputes arising from the third party machines/software.

While operational errors (e.g. incorrect orders or spills) may be less likely with AI than with human staff, such incidents are usually low in value. The more significant shift is from employment-related disputes to commercial disputes involving third-party hardware and software providers.

Employment claims might be relatively common to some businesses, no matter how well managed they are, and whilst the impact on a business may be significant in terms of the time and resource in managing such claims, the financial impact is not always the biggest threat. By contrast, issues such as system failures, software bugs, or cyberattacks may be less frequent but could pose a far higher risk in terms of quantum and impact. A single point of failure could disrupt the entire business, a level of exposure rarely seen in traditional HR disputes. What poses an even greater risk is that these commercial risks are at the mercy of third parties and no matter how smoothly your business operates and what processes you have in place, you could still be exposed to these issues that are completely out of your control.

On the whole, you would expect the commercial issues to occur less frequently than HR issues, but the impact of them when they do happen could be far more significant and potentially catastrophic. It still pays to be as prepared as possible and it is really important to make sure you understand the nature of the commercial agreements you are entering into such as; what the respective obligations of you and the supplier are; what happens when the software goes wrong and who is responsible for fixing that; what happens to your lost profits during any down time; what happens if you just can’t agree and need a resolution promptly and what happens if you just want to bring that agreement to an end.

Discrimination cases

The risk of discrimination claims from the public (not employment claims) could also increase. For example, if the AI technology speech says something to a customer that could be considered discriminatory, or if it is hacked to use discriminatory language.

In other parts of the world, we have also seen cases of similar technologies perhaps not recognising certain demographics. While humans are more likely to be aware of sensitivities, the same might not be able to be said for robots. However, for employment claims, if the law were to change in the future regarding harassment (particularly if liability were to extend to situations where a customer harasses an employee) then replacing employees with robots that interact only with third parties could reduce this legal risk for employers. This is a potential advantage for F&B businesses, which are predominantly customer-facing, should employer liability be expanded in this way.

Health and safety

Moving parts such as automated arms could pose an entirely new health risk to both customers and employees. For example, communicating with an employee to ask them to hold off serving a product for an extra few seconds may be a lot easier than a robot. It’s important to check that you have appropriate insurance cover for issues such as this.

Again, the odds of this happening might be slim, but it would present a whole new legal aspect to consider, should such an issue arise.

Product liability

Liability for errors in products served by a robot, such as incorrect orders or ingredients, generally lies with the business under consumer protection and negligence laws. If harm results from the error for example from allergens, then the business remains accountable, although responsibility may shift to the robot or software manufacturer if the issue stems from a product defect. The use of AI does not absolve the business of its legal obligations; it must ensure the technology is properly maintained, tested, and monitored and that it protects itself with the appropriate contractual or insurance cover in relation to its suppliers.

Many of the manufacturers of such software solutions and machines estimate that their machines have around a 98% accuracy rate, however, should the 2% of cases lead to customers with allergies receiving the wrong orders, it could have huge legal implications for the operators.

The verdict?

This isn’t a commentary on the moral and ethical issues that arise from the progression of robots and AI in
our society, but it does raise some fascinating questions. Haven’t we been here before with previous tech
revolutions? The fear about the introduction of computers in the 80’s and 90’s, the internet in 90’s and 00’s
and then more recently, smartphones, all of which raised perfectly valid points around displacement of
jobs, but all of which created whole new industries and opportunities.

However, as long as the coffee is strong and good – and the above-mentioned points are in place – we can’t
wait to try it out!