How will we teach the robots to behave themselves?

The era of artificial intelligence (AI) is upon us. On one hand it is heralded as the technology that will reshape society, making many of our occupations redundant.

On the other, it’s talked about as the solution that will unlock an unfathomable level of processing efficiency, giving rise to widespread societal benefits and enhanced intellectual opportunity for our workforce.

Either way, one thing is clear – AI has an ability to deliver insights and knowledge at a velocity that would be impossible for humans to match and it’s altering the fabric of our societies.


The impact that comes with this wave of change is remarkable. For example, IBM Watson has been used for early detection of melanoma, something very close to home considering Australians and New Zealanders have the highest rates of skin cancer in the world. Watson’s diagnostic capacity exceeds that of most (if not all) human doctors.

Technologists in the AI space around the world are breaking new ground weekly – that is an exciting testament to humankind’s ability. In addition to advancements in healthcare, 2018 included milestones in AI being used for autonomous vehicles, with the Australian government announcing the creation of a new national office for future transport technologies in October.

However, the power to innovate creates proportionately equal risk and opportunity – technology with the power to do good can, in almost every case, be applied for bad. And in 2019 we must move this conversation from an interesting dinner-party conversation to a central debate in businesses, government and society.

AI is a major area of ethical risk. It is being driven by technological design processes that are mostly void of robust ethical consideration – a concern that should be the top of the agenda for all of us. When technical mastery of any kind is divorced from ethical restraint the result is tyranny.

The knowledge that’s generated by AI will only ever be the cold logic of the machine. It lacks the nuanced judgment that humans have. Unless AI’s great processing power is met and matched with an equal degree of ethical restraint, the good it creates is not only lost but potentially damaging.The lesson we need to learn is this: just because we can do something doesn’t mean that we should.

Ethical knowledge

As citizens, our priority must be to ensure that AI works in the interests of the many rather than the few.

Currently, we’re naively assuming that the AI coders and developers have the ethical knowledge, understanding and skills to navigate the challenges that their technological innovations create.

In these circumstances, sound ethical judgment is just as important a skill as the ability to code effectively. It is a skill that must be learned, practised and deployed. Yet, very little ethical progress or development has been made in the curriculum to inform the design and development of AI.

This is a “human challenge” not a “technology challenge”. The role of people is only becoming more important in the era of AI. We must invest in teaching ethics as applied to technological innovation.

Building a platform of trust

In Australia, trust is at an all-time low because the ethical infrastructure of our society is largely damaged – from politics to sport to religious institutions to business. Trust is created when values and principles are explicitly integrated into the foundations of what is being designed and built. Whatever AI solution is developed and deployed, ethics must be at the core – consciously built into the solutions themselves, not added as an afterthought.

Creating an ethical technologically advanced culture requires proactive and intentional collaboration from those who participate in society: academia, businesses and governments. Although we’re seeing some positive early signs, such as the forums that IBM is creating to bring stakeholders from these communities together to debate and collaborate on this issue, we need much more of the same – all driven by an increased sense of urgency.

To ensure responsible stewardship is at the centre of the AI era, we need to deploy a framework that encourages creativity and supports innovation, while holding people accountable.

This story first appeared on Australian Financial Review – republished with permission.