AI is a hot topic for food and drink
Artificial Intelligence can have many benefits but retailers have to be aware of potential risks, says Linzi Penman
Artificial intelligence (AI) is a hot topic. Interestingly, there is no single consolidated definition of AI, and as a label, it is used imprecisely. For those that fear we are at the stage of robots genuinely thinking for themselves, we aren’t – but there has been a significant shift beyond simple algorithms i.e. computer software programmed to to identify patterns and automate repetitive functions.
Most of us now encounter AI on a daily basis without noticing it. Social media feeds which show you the posts you are likely to engage with, music streaming platforms which suggest new music you may enjoy listening to, or chatbots which help renew insurance policies are all using a form of AI.
We are now seeing what is commonly defined as “weak AI”; systems programmed with algorithms to reach conclusions and predict future behaviour by learning from data patterns. The more data fed to the system, the more accurate the system becomes in predicting future behaviours.
The scottish government has flagged the food and drink industry, worth around £14 billion each year, as a key growth sector in its economic strategy. Particular emphasis will be placed on helping the sector recover properly from the pandemic once life returns to normal. With this in mind, how can AI benefit the industry?
AI has the potential to be harnessed in a variety of ways, including recruitment decisions, chatbots and managing environmental impacts by cutting down waste (which, according to the Food and Agriculture Organisation, is approximately 1.3 billion tonnes globally every year).
Algorithms that accurately predict sales based on data inputs, such as weather and traffic, mean retailers can better connect supply to demand and ensure they don’t order surplus. An “Ai-bin” has even been developed in the UK, using a camera and smart scales to identify the types of food commonly discarded; allowing retailers to quickly change offerings based on trends, such as the unprecedented rise of Veganuary.
With a more focused revenue drive lens on, retailers can use AI to analyse the vast amounts of sales data collected via smart order kiosks or from offering free wifi services and use this data to personalise the customer experience (e.g. tailored menu items). But to fully benefit from the potential advantages of implementing AI, retailers have to be aware of the potential risks in order to mitigate them.
The UK’S data protection regulator has recently developed an auditing framework for AI. There are some fundamental principles for retailers to consider from this framework.
Firstly, does the AI system have an ethical purpose, respecting fundamental rights and core values? for instance will it be used to determine potential audience segments to target personalised marketing offers based on location tracking and preferences – if so, are there processes to ensure that this complies with data privacy requirements?can steps be taken to align with changing regulations?
Secondly, retailers must consider if the AI system is trustworthy and transparent in its decision-making. Is the underlying technology robust and reliable? For instance, will it be used to flag pre-determined ingredients connected to allergens, and if so, are there guarantees in terms of the accuracy, and can this be audited? Can the AI solution adapt to an unexpected surge in foot fall? what happens if it malfunctions and there is insufficient stock or the system inadvertently disc loses personal data of a customer?
There is no simple one-size-fits-all solution. The negotiations with any AI provider and consideration of data concerns therefore deserve careful attention on a case-by-case basis.
Data analytics enable restaurants to
gain valuable insights from their information which, in conjunction with AI, can predict trends, increase sales revenue, reduce operational costs and drive more sustainable choices. but, to maximise these benefits, businesses must ensure AI complies with applicable ethical and regulatory standards and that sufficient contractual governance is in place to manage any risks.
Linzi Penman is a Senior Associate, DLA Piper