GeekSpeak – how do we prevent an AI from going rogue?

When more and more AIs are interacting with humans and becoming powerful there will be many cases where we wish there was an off switch.  Indeed it is already happening.  Microsoft had to delete its chatbot Tay because it became racist after talking to people on Twitter.

How can we prevent an AI from turning against us?  Should we make laws?  Will that be enough? Who will be responsible for the racism or sexism of AIs? Can AIs be controlled by law?

Maybe we simply have to accept that AIs will be independent.  Some will be good, some bad.  Some will not understand the distinction.  Will we all use AI protection in the future as we use virus scanners now?  Can we trust good AIs to protect us against bad AIs?

Or must we accept that we are heading for digital oblivion?

How can we escape our future calculators?  Come and discuss our survival at GeekSpeak.

IM Vulcan Viper if you have ideas for future subjects.

http://maps.secondlife.com/secondlife/Danmu/143/127/51
________________________________________________


Best wishes and have a very happy Christmas

Kathen

Reactions: