IN THIS CHAPTER

  • Loss of control

  • Corrigibility

  • Illusion of control

do we need a killswitch for AI?

For as long as humans have dreamt about artificial intelligence, they have also feared losing control of it. 

Even Alan Turing suggested “turning off the power at strategic moments” to maintain control.  

But what Turing didn’t count on was that a truly intelligent machine would realise it couldn’t achieve its objectives if it could be switched off, so it would take preemptive steps to prevent that.