Six ways to build robots that do humans no harm

robots and other automated systems are getting ever smarter. At the same time they are also being given greater responsibilities

But should they be trusted to take on such tasks, and how can we be sure that they never take a decision that could cause unintended harm?

Keep them in low-risk situations

Likelihood of success: Extremely low. Engineers are already building computers and robotic systems whose actions they cannot always predict.

Do not give them weapons

Likelihood of success: Too late.

Give them rules like Asimov's 'Three Laws of Robotics'

Likelihood of success: Moderate

limits of any rule-based morality

Program robots with principles

Likelihood of success: Moderate

Educate robots like children

Likelihood of success: Promising

learn as they "grow up"

Make machines master emotion

Likelihood of success: Developing emotionally sensitive robots would certainly help