“AI AGENTS MAY “LEARN IN THE LONG RUN TO AVOID SUCH INTERRUPTIONS … WHICH IS AN UNDESIRABLE OUTCOME.”
At first I was totally okay with a “big red button” to disable a rogue AI that was either a danger to itself or, ahem, humans. The researchers noted that it’s possible the intelligence could eventually learn to avoid that situation by disabling it’s own shutdown mechanism….so I have some more thinking to do on the “big red button” issue before weighing in.
Source: Researchers want a ‘big red button’ for shutting down a rogue artificial intelligence | The Verge