When I purchased my smart speaker back in 2014, a colleague and I were keen to test how “intelligent” its personal assistant functionality really was. So we asked it a series of related questions about one of my favorite actors, Cate Blanchett. And while we found that it had no trouble with direct questions such as “How old is Cate Blanchett?” and “Where was she born?” (Australia, of course), it couldn’t tell us how old she’d be in 2020.
Such are the present limitations of commercially available consumer-facing artificial intelligence (AI) assistants. Currently these assistants are capable, based on a verbal request, of translating, interpreting and mining structured data. But they still lack the ability to truly “think.” So for now at least, artificial intelligence plus human intelligence equals optimal performance.
Testing the limits of AI technologies
This is a critical issue for organizations to recognize, understand and balance. AI has the potential to make them more efficient. It could also support better decision-making based on big data and the targeting of services and offerings to better align with their customer base.
But to make the most of these opportunities — and more pointedly, to mitigate the associated risks — organizations need to understand the boundaries and limitations of the AI technologies they adopt. And it’s the board’s responsibility to ensure they do it.
The challenge for boards is therefore to make sure they ask the right questions of management.