There is an old consulting adage that is often applied when new technologies come along: ‘I’ve got a hammer, so the answer to your problems must be a nail’. This is particularly true of Artificial Intelligence (AI) right now. Many enterprises are asking ‘how can we use AI in our business?’ (or, even more flippantly, ‘can we have an AI?’).
Whether this is the consultants selling the idea that AI is the silver bullet solution to all their problems, or the enterprises themselves being taken in by all the hype and froth around AI at the moment, one thing is for certain: this is the wrong question to ask.
So, let’s put the hammer away for a while and turn that question around. Let’s ask, ‘Can AI address some of my business objectives and challenges?’. Of course, if you don’t understand or appreciate what your business objectives and challenges are, then you have far greater problems than any technology can address. But let’s assume these are known and start to think about how we might answer that question.
To do this we need to think holistically and collectively about the possible solutions to realizing those objectives and meeting those challenges, which will undoubtedly involve more than just AI (perhaps a saw, a chisel, and a drill as well as our trusty hammer). And we will also need to understand how we get to those solutions, which will need more than just data scientists and developers.
We must always bear in mind what we are trying to achieve, with AI as a potential way to enable the solution, but not necessarily the overall solution itself.
Geoff Mulgan is the CEO of the London-based innovation foundation Nesta, and in an article, he wrote recently he makes a similar plea by wanting to put intelligence as the outcome rather than the input. Many of you will be aware that AI inherently looks at very narrow applications because each AI can do one thing, and one thing only, very well, such as text classification, prediction or image recognition (even to the point that an AI trained to recognize pictures of dogs, for example, wouldn’t be able to recognize pictures of cats). So, the natural approach of any AI technologist or AI app developer is to focus on single solutions. But anyone who knows anything about Systems Thinking will appreciate that a benefit in one area usually means a compromise or deficiency in another. To use a slightly contentious example, Facebook has spent a long time optimizing click-throughs on ads but has tended to neglect the privacy of its users.
What Mulgen advocates to mitigate these natural tendencies is a much greater collaboration between different skill sets on the project, something he terms Collective Intelligence. As well as technologists, this could include process designers, psychologists, organizational designers, economists, decision scientists, and sociologists. And, as well as taking in advice from a wider team, we also need to consider all the other technologies that might support the AI (such as RPA, workflow, and portals), as well as how the solution will be impacted or enhanced by rules, norms and complementary innovations in the environment that it is being implemented (data confidentiality has different levels of importance to people from different parts of the world or from different generations, for example).
As the article concludes, this whole idea of Collective Intelligence, of which AI is just a part, “forces you to address the different contributions of machines and people to observation, creativity, memory, and judgment. It takes you quickly to combinations and hybrids. And it encourages humility on the part of the people who come from particular disciplines and backgrounds”. As well as being a great way to look at how to break through the AI ‘hammer and nail’ issue, it may also just be the best way to look at everything we do in our technology-heavy lives.