Safeguarding Against Artificial Intelligence
Author of “Superintelligence,” Nick Bostrom, a Swedish philosopher at the University of Oxford and founder of the Institute for Ethics and Emerging Technologies says that “when A.I. teaches itself how to learn without supervision, it will be able to learn so fast that it would overrule humanity and take over.”, a concept first coined as ‘Singularity’ by mathematician Stanislaw Ulam 1958.
Now, the scientists who build the A.I. algorithms need to install an “interruption policy” that would “forcibly change the behavior of the agent itself.” This signal would convince the machine, that at this point would be ignoring human commands, that it should stop and Google is planning on doing just that.
Google DeepMind is taking steps, in conjunction with The Future of Humanity Institute, to outline a framework for stopping out-of-control A.I.
A paper called “Safely Interruptible Agents, states “Safe interruptibility can be useful to take control of a robot that is misbehaving and may lead to irreversible consequences. If such an agent is operating in real-time under human supervision, now and then it may be necessary for a human operator to press the big red button to prevent the agent from continuing a harmful sequence of actions – harmful either for the agent or for the environment – and lead the agent into a safer situation”