Are ‘mind children’ the future of reproduction?

7 hours ago 14

A few months ago, an AI researcher from Europe attended a dinner party in Silicon Valley. During one of the many courses, the host addressed his guests, all of whom worked in AI. The researcher paraphrased his message like this: “Isn’t it amazing that we are the last generation of humans who will need to think about procreating biologically? We were lucky enough to be born at a time where we can simply upload our consciousnesses instead.”

“I didn’t see that coming,” the researcher told me. “I was just enjoying my fish.”

But the host was serious. His words struck the researcher as the kind of comment a well-informed person might have made 100 years ago, once antibiotics had been invented: “Aren’t we lucky that we came after?”

Suddenly all the guests were talking about “mind children”, and the researcher turned to their neighbour to ask what this phrase meant. “He said, ‘Oh this is the book,’ and, ‘Haven’t you read the book?’ and, ‘Oh my God, you should really read the book.’”

The book in question was Hans Moravec’s Mind Children: The Future of Robot and Human Intelligence, which was first published in 1988, and which at the time, according to economist and futurist Robin Hanson of George Mason University, caused a big splash in a small pond – the community of robotics and machine-learning experts to which Moravec belonged.

Moravec’s book is more philosophical treatise than technological manual, but the central idea is that cultural evolution has long since taken over from biological evolution as the most powerful force shaping humanity, and the logical extrapolation of this is that the information that encodes our future selves would soon be packed into hardware and software rather than DNA. These mind children could be equipped with soft, squishy bodies, like real children, but they could also take a kaleidoscope of other physical – or indeed non-physical – forms.

Moravec observed that the ultimate consequences of this revolution were unknown, but he also appeared to welcome it. Within a century, he wrote, machines would exist “in whom we can take pride when they refer to themselves as our descendants”.

Hanson shares his conviction that the revolution is inevitable, as soon as AI attains something experts agree to call human-level intelligence. “We are going to generate an explosion of things like us in the future, who will be different from us in many ways,” Hanson says. “To the extent that they have minds somewhat like ours, they are our mind children.”

Angela Aristidou, who studies the real-life deployment of AI at University College London, is not surprised that Moravec’s book is enjoying a revival. She says that what in 1988 might have read like science fiction – and still might to most of us – looks eminently realisable to those in the know. Elon Musk’s pronatalist stance is the exception among tech types, she says, while the idea that the clock is ticking on biological reproduction is far more common – and the harbingers of that (perhaps self-fulfilling) prophecy are there for all to see. Delegates to this year’s Nvidia GTC in San Jose, California, a major AI conference, were treated to an AI avatar of Nvidia CEO Jensen Huang, for example.

Then there’s the phenomenon of human-AI weddings. Such unions obviously can’t produce biological offspring, but since the human in the relationship has usually created their ideal romantic partner in the AI, Aristidou asks, rhetorically: “Why wouldn’t they also devise their ideal child?”.

In thinking about this post-biological future, though, we have to stretch our concept of “child”. The new entity could be an AI that human parents lovingly and jointly sculpt to meld the perceived best parts of themselves – as is already technically possible with gene editing in biological reproduction – but given that we’ll be doing away with birth, death and generations, as these concepts are ordinarily understood, it could also be something quite different.

A human could simply upload their own consciousness so that it outlives their physical shell, in which case the child is something closer to a clone. The human could transfer some of their consciousness into their AI companion, or conversely devise an AI companion that they perceived to be the opposite of themselves, in the belief that opposites attract. In all cases, a new entity emerges, but the line between self, partner and offspring is blurred. If that sounds incestuous, remember that there’s no risk of the medical conditions associated with inbreeding, though there could be others.

Aristidou doesn’t doubt that AIs can augment human relationships. They have been shown to be helpful as assistants in a therapeutic context, for example, or in overcoming loneliness. But she’s concerned about what happens when AIs become human substitutes. If a human can delete their AI spouse, she says: “How does that work as an equitable marriage the way we understand it?”

She also worries that a two-tier society will emerge, in which a tech-savvy, well-resourced elite customises its AI creations for high realism, maintaining control over their settings and updates, while everybody else has to make do with cheaper, off-the-shelf products that place them at the whim of developers – “as if there’s three entities in this relationship: the human, the AI companion, the AI developer”. Among the many ethical, legal and practical issues this throws up is whether the developer would be considered a co-parent to a mind child.

Hanson says there are legal scholars and ethicists thinking about such things, but until society takes its postbiological future seriously, the safeguards they are proposing have no hope of being debated, let alone implemented.

Nor is anybody discussing arguably the thorniest issue of all: is humanity staring down its own last act? Hanson says that the emergence of more complex lifeforms doesn’t necessitate the extinction of older, simpler ones – or there’d be no more bacteria on Earth. But if that thought strikes you as less than reassuring, take a leaf out of Moravec’s book and focus on the positives. “Very little need be lost in this passing of the torch,” he wrote in 1988. “[It] will be in our artificial offspring’s power, and to their benefit, to remember almost everything about us, even perhaps, the detailed workings of individual human minds.”