Business AI needs to focus on psychology, not technology

There’s no doubt now that AI works, and it can work very well indeed – as a technology, at least. And as one of our recent research projects reported, AI-based machine learning (ML) is increasingly being used to automate (or part-automate) complex processes such as logistics planning and IT network monitoring. It is even being pre-integrated into industrial machinery and the like.

These applications are typically fairly narrow though, and while they can be transformational within their specific niches, they affect relatively few users. Of course, ML of this kind still needs domain expertise and analysis – to ensure we’re not automating faulty assumptions, for example – but it tends to be designed to assist or enhance an existing role or process.

Where it gets interesting – and not a little worrying, too – is when AI moves into the wider world of business transformation. It’s important because for many organisations, applying AI to general business workflows and processes could represent a new stage in the process of digital transformation, and one that has the potential to be disruptive in all the wrong ways if it’s not managed well.

The desire to promote AI-driven business transformation is a large part of why the big IT vendors are paying a lot more attention now to the human domain – to the ‘softer’ sciences and skills, such as business psychology and change management.

Successful business AI needs trust and transformation

By their very nature, broader AI solutions need to be scaled out across an organisation in order to be effective, and that in turn requires transformation and culture change. The AI needs to be integrated within the business culture – and as part of that, it needs to be trusted by the people who must work with it and will be affected by it. That of course includes staff trusting it to help them work better and smarter, and to do more with less, not simply to put them out of a job!

This is why the thought-leaders and IT suppliers that understand this area best are talking more and more of concepts such as AI ethics, human-centric AI, and ‘AI as human augmentation’. Those concepts are taking a long time to trickle down into end-user organisations, however.

For example, at Microsoft’s Future Decoded event in London last year, it presented findings from a study by researchers at London’s Goldsmiths College on the state of AI in the UK. Among the results highlighted were that 96% of people said they have never been consulted by their boss on the introduction of AI, and that conversely, 83% of business leaders said their employees have never asked about AI.

You might ask why this should matter, when it is hardly routine for bosses to consult their staff about introducing process automation, or workflow management? Quite simply, it matters because of that transformational and cultural aspect. We’ve written before about the importance of the human factor to the success – or otherwise – of workplace transformation, and AI makes it even more important.

How do you learn what to do when AI gets it wrong?

After all, without those consultations, how can an organisation scale up its use of AI beyond niche and pilot projects? How can it be confident about spotting and eliminating the kind of AI bias that’s been in the news in recent months, and which has the potential to completely derail an AI project? How can it acquire and develop the skills not just to implement AI, but to know what to do when AI gets it wrong?

And, routing back to those soft skills, how can the organisation scale AI without getting its people on-board – without getting them to understand and accept what integrating AI into their working practices will need, and of course how they will benefit from it?

This kind of thing – how to get buy-in for change, and the need for change to benefit the individuals affected by it, as well as the organisation’s bottom line – is well understood in business change and re-engineering circles. The challenge now in AI is for the technology to take a back seat. It’s time to focus instead on how to integrate it into business strategy and culture – and yes, into business governance, ethics and responsibility.

CIO
Security
Networking
Data Center
Data Management
Close