What are the big themes in machine learning and how do they affect businesses?

Transparency

Perhaps the greatest short-term concern about AI is the simple fact that it takes decision making out of human hands. This can potentially have enormous consequences if Machine Learning algorithms have learnt the wrong lessons (for example, acting in a way that might be seen as being prejudicial). To police this, it’s crucial for AI professionals to have visibility into how AI reaches the decisions that it does and the ability to identify red flags.

What’s more, when this overlaps with regulation (for example, forbidding discrimination in hiring or the provision of financial services) it may be necessary to extend this transparency to government and, potentially, to the individuals being analysed as well. This opens up a secondary concern, however. While legislation like GDPR stipulates precisely how personal data is used, AI may be endlessly accessing data and segmenting and collating user profiles. 

Given these issues, there are increasing calls for regulation. In the meantime, the industry is filling the gap with organisations like OpenAI. Google recently set out to create an AI ethics board, but it quickly collapsed in the midst of controversy relating to the individuals who had been picked for the programme. It seems likely in the near future, that businesses will need to be able to demonstrate accountability and responsibility when it comes to AI governance standards.

It’s important to remember, though, that AI is not a monolithic thing. Instead, it’s a collection of different technologies, at different stages of development which we will interact with in a variety of different ways – and in many cases, we’ll be unaware of its guiding influence.

Looking ahead

While there are certainly justified concerns about AI, the sheer versatility of the technology and the enormous benefits that it offers to businesses and consumers will see it becoming enmeshed in every part of our lives. A key question for the future, though, is whether the key driver for advances lies in engineering talent or access to data.

Incumbents with large-scale, organised and integrated datasets will be at a distinct advantage. However, China’s new social credit system – which tracks behaviour, interactions and relationships, online and offline – illustrates the potential to create truly vast databases and to integrate these with real people, places and property. It may just be that this data-rich environment gives Chinese businesses a critical advantage in the AI arms race in the years ahead.