Short Bytes: Just like Stephen Hawking and Elon Musk, do you fear that AI could summon “demons” and endanger our existence? To eradicate these possibilities, Google and Oxford have partnered to develop a “big red button” that can terminate any unwanted action of an AI agent.
Google’s AI agent has already beaten humans at various games like Go and similar progress is being observed in other areas.
Now, researchers from Google and Oxford have partnered to mitigate the risks that an uncontrollable AI could pose. The findings of the research are published by Oxford’s Future of Humanity Institute under the name “Safely Interruptible Agents.”
Google’s DeepMind is one of the most advanced AIs out there and Google is making sure that humans remain the ultimate controllers of its robots.
In their paper, the researchers explain that AI agents are “unlikely to behave optimally all the time.” As a solution, they have talked about a “big red button” to prevent AI agents from creating some harmful situation.
The researchers involved in this collaboration claim to have created a framework that will give a human employee the power to interrupt an AI and make amends or stop it.
The authors of the paper write that some AI algorithms like “Q-learning” are already interruptible and they are looking forward to bring the same functionality to other areas.
If you are willing to know more about this research on AI, here the link to the original paper.
Did you find this article helpful? Don’t forget to drop your feedback in the comments section below.