Will we be able to teach ethics to robots?

The emergence of 5G technology combined with the fast-pace machine learning evolution, as well as all the tools that raises Artificial Intelligence possibilities, are enabling what specialists have been calling the fourth industrial revolution, which is definitely going to lead us right to a still unreliable robot era.

While the IT industry has been focused on launching high engineering advances that will provide autonomy to machines and popularize driverless cars, delivery services by drones and other kinds of specific knowledges that will be perfectly filled by robots, developers and society face a tough question not yet solved by machine learning: how to program ethic values to artificial intelligence?

Will a robot be able to make decisions beyond its algorithms, driving to the right way when facing a dilemma? How are human scientists going to create a moral statement, or some standards simulator, that contemplates so many unpredictable situations and which allows the machine to decide ethically?

This matter clusters in a big question specialists from philosophy, psychology, laws, besides governments and of course computer science. Imagine machines at the service of army, police, traffic, education, healthcare and other functions sensitive to ethical rules.

Unesco delivered important contribution in terms of proposing alternatives to fill this gap. In 2017, its Commission on Ethics of Scientific Knowledge and Technology reported the concern regarding the technic features of robot development, as the main point. The commission aimed the different between cognitive machines – from which actions are based on previous experiences – and deterministic robots – that have behavior modulated by a programme. Considering the arguments, Unicef proposed recommendations.

The Royal Society, a London non-governmental institution in charge to promote scientific knowledge, suggests an Ethical Governance, defined by “a set of processes, procedures, cultures and values designed to ensure the highest standards of behavior”.

Technology advances run faster than the rise of a consensus. But we don’t seem to be prepared to go forward to robot’s era while we Artificial Intelligence keeps advancing in its narrow capabilities, and not in a general way. I mean, AI keeps growing faster but to respond to specific demands, very strict tasks, pragmatically. The biggest deal will be to transform AI to make machines act like humans.

The fact is that all the planet must be involved in this discussion in order to state a single regulation, as a pact, allowing the inevitable revolution ahead.