To quote Dad’s Army’s Private Fraser – “We’re doomed!” And the cause of humanity’s downfall? Well if you believe pronouncements late in 2014 from the likes of Professor Stephen Hawking and PayPal founder and billionaire Elon Musk, the latest existential threat to the species is runaway and uncontrollable Artificial Intelligence (AI).
Achieving the kind of ‘superhuman’ AI which alarms Hawking and Musk remains some years off. But of late, AI has been enjoying something of a renaissance with exciting breakthroughs in domains such as machine learning, computer vision and image recognition, and hardware advances such as the neuromorphic processors pioneered by IBM and Qualcomm, and the massive processing potential afforded by supercomputers and the cloud.
We have also been hearing a great deal about automation and the threat posed for many established job roles and types of work. I discussed this in a previous Engineering & Technology magazine interview. There are many who believe the changes ushered in over the coming years by increasingly capable AI will dwarf the transformation and upheaval of the transition, from an agricultural to a mechanised economy brought about by the Industrial Revolution.
All this talk of threat to species, massive transformation and revolution can perhaps make AI see a little intangible and remote to senior decision makers in organisations, but I believe AI should on the agenda of management boards today.
AI is already here and very real. Many companies have been using it for years but may not be aware they have been doing so – the Bayesian algorithms that underpin card fraud detection and engine management are examples of AI in action. The kinds of AI that can be bought today are perhaps more overt and their effects more wide-ranging.
AI’s ability to significantly impact an organisations’ performance is increasing. Using genetic algorithms that optimise towards particular goals, AI can come up with product designs that would be difficult, if not impossible, for people to replicate. AI is becoming increasingly adept at handling natural language interactions with people in a customer service environment.
As we see with IBM’s Watson in its application supporting oncologists, learning AI can act as trusted adviser that supports people in their decision making. If an AI can help a clinician diagnose cancer, how long before AI will routinely support boards and senior management in their own decision making?
It is not just the performance and operational efficiency improvements of AI that should interest CXOs. AI has the potential to affect brand and customer perception in unpredictable ways. Loss of jobs due to automation can be controversial, and the willingness of both customers and employees to deal with automated intelligent systems is not uniform. Indeed it can vary widely and ignite strong feelings and passions.
Personally, I am glad to see the likes of Hawking and Musk pointing to the potential future hazards of AI. There are undoubtedly equally significant benefits for humanity and these need to be thoroughly explored and safeguarded. The ethical, moral and philosophical debate is complex but needs to be had – and we should be having it now.
Similarly, I think it is important that company boards lead their organisations in a debate on how to approach AI. This will help them to understand and keep pace with what’s possible in AI, anticipate what’s coming next and to decide how automation and impacts are best managed across people and operations.
We’ll be exploring more about AI on the blog this year, so watch this space!