Forget evil AI – the real risk is far more insidious | The Guardian

It’s easy to get carried away with dystopian visions of sentient machines that rebel against their human creators. However, the real risk posed by AI – at least in the near term – is much more insidious.

It’s far more likely that robots would inadvertently harm or frustrate humans while carrying out our orders than they would become conscious and rise up against us. In recognition of this, the University of California, Berkeley has this week launched a center to focus on building people-pleasing AIs.

The Center for Human-Compatible Artificial Intelligence is lead by computer science professor and artificial intelligence pioneer Stuart Russell.

“The risk doesn’t come from machines suddenly developing spontaneous malevolent consciousness,” he said. “It’s important that we’re not trying to prevent that from happening.”

Read more: The rise of robots: forget evil AI – the real risk is far more insidious | Technology | The Guardian

Don’t forget to share this via , , Google+, Pinterest, LinkedIn, Buffer, , Tumblr, Reddit, StumbleUpon and Delicious.

Mike Rawson

Mike Rawson has recently re-awoken a long-standing interest in robots and our automated future. He lives in London with a single android - a temperamental vacuum cleaner - but is looking forward to getting more cyborgs soon.

Leave a Reply

Your email address will not be published. Required fields are marked *

Forget evil AI – the real risk is far more insidious | The Guardian…

by Mike Rawson time to read: 1 min
Hi there - can I help you with anything?
[Subscribe here]
 
More in Machine Learning, News
Sewage robot
Luigi The PoopBot Wants To Know What You’re Ingesting

Over the last few years, scientists have turned to a new source to discover statistics about everything from diet to...

Close