The mission, should you choose to accept it, is:
'What simple rule or rules would you write your AI algorithm with to avoid harm coming to humans?'
Your rules, nobody else.
Of course, AI was out of control when it was first conceived but if the simple rules are followed for a while maybe they'll become standard.
The mission, should you choose to accept it, is:
"Can you define AI in terms that make sense from a standpoint of computer science, rather than Luddite fearmongering?" Because then we can talk about how unhelpful articles like this are.