Put yourself in his shoes.  In charge of a company that stands at the forefront of an area of technology with the potential to change so much of what we know as the norm in ways that perhaps we don't fully appreciate. 

Excited? Terrified?

The thing with AI (or as this article refers to, Artificial General Intelligence) is that it has the potential to do an awful lot of good.  Yet in a pessimistic world (and we lawyers can sometimes be accused of being glass half empty I'll grant you) that sometimes gets lost in the debate.  However (he says, draining his glass) we all know that with every upside there's the inevitable balance.  In this case they're the stuff of literature masterpieces and Hollywood blockbusters.

He knows he needs to control them.  We know we need him to control them.  But the moral debate is wider than that.  Our societies, for there are many, are each shaped by historical events, cultural development in isolation that only now, in this modern interconnected world, are we really understanding and truly experiencing the differences.  

So who decides the moral framework and parameters that should shape the world that AGI develops in?  Who decides what is right, and more importantly perhaps, what is wrong?

AI is moving at a phenomenal, unexpected pace at the moment.  The challenge is to ensure that in a bubble of competing development the human mind is not lost to the opportunity and stays alive to the risks.  At a very simplistic level, we've all experienced the feeling when a machine is going out of control and someone just unplugs it....relief.  Phew.

I've written previously about the need for the law to adapt as AI development accelerates here: http://mediatech.footanstey.com/post/102df8f/adapting-the-law-for-an-ai-world and that's true.  But maybe society needs to move too - this article touches on the notion that the understanding of many may focus more on the negative and less on the benefits.

History tells of stories of other scientific minds who later regretted their creations.  The goal for AI is surely, simply, to ensure we do just that so that we don't look back with similar regret.