A couple of weeks ago I was on holiday, standing on a windswept Cornish golf course, firmly intending not to think about work. But then a dog-walker appeared, ambling along a footpath, and I found myself pondering parallels between dogs and AI.

By domesticating wolves, early humans gave themselves a productivity-enhancing assistant. Dogs’ superior noses and ears complemented human hands and brains, making Homo sapiens a better hunter. Some people even grew to love their canine companions. Yet it was far from obvious, back when dog technology was first being developed, that it was safe. We were literally inviting wolves into our caves. So the big question about AI is a bit like the one our ancestors presumably asked about dogs. Will this awesome helper serve us faithfully, or slip the leash and bite us? Our cover package this week chews it over.

In an extended Briefing, we look at the evidence as to whether AIs could one day become conscious. Anthropic’s researchers have discovered something that looks like an internal thought process in its Claude model, which is related to its outputs but normally invisible to human users. This is probably not yet consciousness, but may be a step towards it. Meanwhile an Australian firm is putting real human brain cells, grown from stem cells, onto silicon. No one is sure where this may lead.

Sooner than many people expect, we may have to grapple with the possibility of algorithms waking up and feeling. Already, many users treat their AI chatbots as if they were sentient. And Argentina’s president is proposing that AIs be allowed to run corporations—a limited form of legal personhood. In a leader, we argue that humanity should never give rights to AIs. If they become superintelligent, but have a right not to be switched off, they may end up as our masters. And we have no idea how AIs would treat their pets.

On the Insider our editor-in-chief, Zanny Minton Beddoes, discusses AI and other topics with Yuval Noah Harari, one of the world’s more pessimistic superstar thinkers. It’s not one to miss; the episode is live now. If you are optimistic about AI—and up for some experimentation—I’d encourage you to sign up for The Economist for You. Our personalised newsletter launches tomorrow and uses an in-house AI model to comb through our coverage to build a reading list tailored to your tastes.