Machine learning, process automation, affective computing and sentiment analysis, and natural language processing (commonly collectively referred to as 'Artificial Intelligence' or 'AI'), are transforming the way we live and work.

The trend towards the digitisation of business that had its roots in the middle of last century and exploded with the advent of eCommerce and cloud computing in the early 2000s, has inevitably led to two key resources: the availability of incredible processing power; and massive quantities of highly available information. Together with the knowledge to construct algorithms and neural networks to make use of these resources, an AI perfect storm has arisen which will see technological disruption across all sectors. The revolution has already begun.

The world’s largest online-only grocery retailer is using machine learning to improve operational efficiency and customer care. Ocado's contact centre was receiving many thousands of messages daily and Ocado wanted a better way of prioritising them for response. It wanted to respond faster to time-sensitive messages (such as a customer who is missing an item from their order) than to those where a small delay wouldn't matter (such as a customer sending a thank you message for great service). By implementing an AI solution, Ocado was able to respond to urgent messages 400% faster, free up 7% of customer service agents' time, and predict customer needs to optimise the ordering process. The new system is 80 times faster than the old one, and at only two-thirds the cost. In a sector with small margins, these kinds of improvements provide genuine competitive advantages.
Drug development is a long, laborious, and expensive process; and there are advantages to patients, healthcare providers (including the NHS), and manufacturers in getting products to market safely and efficiently. It typically takes looking at around 2,500 compounds—four or five years of effort—to get to a candidate formula for a new drug. Multinationals like GlaxoSmithKline are using AI to digest huge quantities of data to anticipate how new compounds might interact to vastly reduce the development time for new medicines. Early signs are that the new system could cut this development time by as much as 75%—a massive improvement which will save lives. Separately, researchers have developed an AI for the early detection of heart disease for implementation in NHS hospitals. Estimates suggest this could save the service up to £300 million each year.
Drivers across the UK receive parking tickets that they believe were given incorrectly, but without legal advice many do not know how to go about challenging them. Due to the low value of fines (typically around £80 or less), obtaining professional legal advice was simply not financially viable. That was until Joshua Browder created the 'DoNotPay' chatbot, providing an automated process to advise on appealing parking tickets. Using AI, he made it possible to serve this low-value/high-volume market, and has since expanded to cover a range of legal issues including asylum appeals and consumer law issues. It also serves the public good, by improving access to justice for day-to-day legal problems affecting thousands of people.


Artificial intelligence presents a wide range of opportunities for business—as a back-office tool to drive efficiency, a way of delivering services more efficiently to drive profitability or customer satisfaction, or as part of a new product or offering to generate new revenues. Of course, there will be numerous different legal considerations depending on the purpose of the AI solution. Whether you're the vendor or the customer, the contract will need to reflect your priorities and include a sensible balance of risk and reward.

If you're integrating AI into your existing business, you'll need to consider any points of interaction with legacy systems. If you're offering an AI solution, how you can help your customer de-risk the implementation? Artificial intelligence is unlikely to have been contemplated in contracts for legacy systems, and an important task for the legal team is to look to existing contractual arrangements which may impact or be affected by the implementation of the new solution.

When letting an AI loose on data, any potential cybersecurity, data privacy, and GDPR-compliance aspects need to have been resolved—otherwise your innovative use of customer data might become an expensive white elephant! And with AI learning all about the inner workings of your business, great care will be needed to make sure that information doesn't end up in the wrong hands. Data privacy by design means careful planning when designing the solution, and how it will fit into an existing business environment, not trying to retrofit compliance and security later.

With all of the above in mind, both customers and vendors need to be careful to follow the right processes when contracting for the delivery of a new AI:

  • To the customer—are your technical and business teams signing up for free licences or low cost pilots without legal scrutiny? These can turn out to be Trojan horses, which bring unknown risks within the walls of the customer's organisation.
  • To the vendor—do you fully understand how the customer is going to be using your solution? Could the customer's (lawful) use unintentionally put you at risk of legal or regulatory challenge?

The market for AI solutions is growing rapidly, but is relatively immature. Many organisations are trialling and implementing AI technologies, but few are considering the legal consequences. We've seen that the impact of poor decision making in AI implementations leads to wasted time and money, and potentially serious compliance issues.

Recognised as leaders in our field, and for our highly commercial and pragmatic approach, we've got the expertise and experience to help you achieve your strategic objectives when it comes to artificial intelligence.