By Lillian Barnard
The events of 2020 have opened the world’s eyes to digital transformation as never before. Despite many people not being aware of the term, remote learning, working and socialising has seen people embrace more technology than ever. Much of this adoption is of technology that has been around for quite a while, but more and more, we’re seeing some advanced services and solutions being spoken about as countries and companies look to emerge from the crisis stronger. One of these, probably the most prominent, is AI.
AI seems to make a regular appearance in success stories surrounding COVID-19 specifically. Lebanon is a case in point where citizens now have access to a health bot with an inbuilt symptom checker to conduct self-assessments for the virus. The sophisticated bot is able to identify high-risk patients and direct them on to further treatment and consultation, helping to alleviate the capacity of frontline healthcare workers.
Not surprisingly, advancing AI across the region is as great a priority as it’s ever been. But in order to move forward, we first need to address some of the major stumbling blocks that might hold us back – starting with the regulations that surround AI. While the regulatory landscape might not be the most obvious hurdle to adoption of AI it is the most significant.
A recent Microsoft-EY study shows business leaders view regulatory requirements as the number one risk associated with AI. In fact, almost half of the respondents to the study said they had concerns about adhering to existing regulatory requirements, especially against the backdrop of a regulatory landscape that is still evolving.
Data security in the spotlight
This is, even more, the case as governance has been pushed to the fore by the recent shift to remote working. Organisations are finding it trickier than ever to ensure data security and governance as they try to maintain business continuity amid the crisis. A global study by Harvard Business Review shows that even though 77 per cent of organisations believe an effective security, risk, and compliance strategy is essential for business success, an overwhelming 82 per cent say that securing and governing data is becoming more difficult.
Understandably, leaders are worried about investing in AI while the spotlight is on governance and the rulebook is still being written. It’s a catch-22. While businesses are nervous to expand their use of AI, they also risk being left behind by their industry counterparts if they don’t consider how best to leverage the technology across their operations. Research shows the majority of businesses in MEA are already piloting AI projects. The question is, how best to move forward?
Fine-tuning governance frameworks
While the regulations surrounding AI need to be considered throughout its development cycle, responsible AI typically begins with good data governance. It’s why AI leaders across the region are so focused on developing solid data governance frameworks. Even those with the most ambitious AI agendas are still heavily focused on fine-tuning their infrastructure.
Understanding the value of data sets the right tone
AI frontrunners also know that good data governance begins with ensuring the value of data is understood and prioritised throughout the organisation. Engaging the C-suite in defining a data governance strategy is key to successfully implementing AI.
We see this approach at work in ABInBev, which has moved beyond the early stages of AI implementation and is now starting to roll out machine learning solutions across functional operational domains. Much of the company’s success with AI can be attributed to its senior leadership who have developed a solid understanding of AI. This allows them to deploy the technology across the business’s operations while managing the risks of regulatory compliance and ensuring sound data governance.
Understanding needs to be followed up with accountability. A good way of ensuring data management and governance policies are implemented is to appoint a data steward with a primary focus on ensuring the quality of data. This they should do by monitoring systems for data management irregularities and prioritising those issues with the greatest opportunity costs and business impacts.
When an issue with data quality arises, the data steward can then flag this with senior leaders who are able to prompt teams to scrutinise their procedures and implement remedies to reduce the risk of recurring issues.
When it comes to optimising procedures, one of the biggest hurdles companies face is determining who owns the data, how it is stored, how to access it, and who may have access. This is why implementing role-based access to data is so important. The goal is to provide consistently managed processes around data security and integrity across the designated roles. These steps can help ensure data is accurate, discoverable, accessible and well-managed, ultimately putting organisations on the path to success.
This year’s events have reinforced the idea that AI can significantly assist both business and society when facing challenging times. AI maturity is a major part of ensuring countries across the Middle East and Africa are ready to tackle unforeseen hurdles. Undoubtedly the first stop on this journey is good governance.
Lillian Barnard is the Managing Director of Microsoft South Africa