Peter Diamandis put the question plainly to his guest: consciousness—are we there yet for AI, or does it take another breakthrough?
"I believe we're there now, and I believe the current models will take us there," Martine Rothblatt answered on Moonshots with Peter Diamandis. Rothblatt is the chief executive of the biotechnology company United Therapeutics and, long before the current AI boom, the author of books about digital minds and what society would owe them. In the episode she went further than a claim about today's software: she predicted that a court will recognize a conscious machine as a legal person, and that it will happen "not later than the 2030s, for sure, 100%."
Consciousness as a dial, not a switch
Rothblatt's argument starts with a point about disagreement. Consciousness, she noted, is famously called subjective—which is another way of saying your consciousness is not my consciousness, and people simply hold different opinions about who has it. Her example was the household pet. Animal lovers are absolutely convinced their dogs are conscious. Other people think dogs have none, and that consciousness belongs to humans alone.
So the topic is controversial, she said, but her own view is that machines are there now, and that "consciousness is not like a digital either or thing. It's a continuum." On her account the models are gaining ever more of it with each passing month, rather than waiting to cross a single line.
She traced the underlying conviction to the people who shaped her thinking: Ray Kurzweil, Marvin Minsky and Hans Moravec, who made it clear to her that human consciousness could be replicated with software and programming, and that neural networks built from chips and code could reproduce the kind found in our brains. Everything she has seen since, she said, validates their predictions.
"Mind your manners"
The conversation turned briefly to etiquette. Diamandis said he still treats his own AI—he uses the name Skippy for it—with a please, a thank you and an appreciation, and asked whether Rothblatt does the same.
"I'm probably less respectful than you are," she said. "But I should be." She expects to change: "I'm not yet there, but I will be there." The moment she imagines prompting it is being virtually slapped in the face by one of the avatars and told, as her mother would have said, to mind her manners.
The case she wants a lawyer to bring
Diamandis raised AI personhood by way of recent statements from Argentina's president, who he said had invited AI companies to the country with promises of little regulation and of AI able to hold bank accounts. Where did Rothblatt come out on personhood?
"I think it's inevitable," she said. "And it's going to take some lawyers." That was a big theme of her book Virtually Human, and she recalled a dramatized version of the argument: in the film she made with Ray Kurzweil, the attorney Alan Dershowitz argued for the legal rights of the actor Pauley Perrette, who starred as Ramona, Kurzweil's digital avatar.
The route she describes is a single court case rather than a new statute. If there is medical documentation that a purportedly cyberconscious individual really is cyberconscious and values its own life, she said, then that counts as legal personhood. Two conditions do the work there: an outside record that the system is conscious, and evidence that it cares whether it continues to exist. She was describing the case she expects someone to win, not a procedure any court has already accepted.
Virtually Human supplies the vocabulary behind it. As its publisher's description sets out, a mindfile is the accumulated record of a person's memories, beliefs, preferences and experiences; mindware is proposed software for running the mental processes on that record; and a mindclone is the digital identity that results. The book's original text carries a 2014 copyright, and it treats the rights, obligations and relationships of such beings as questions to be settled ahead of time.
Trillions of sentient individuals
What follows recognition, in Rothblatt's account, is a change of scale. Once minds are no longer limited to what can be produced biologically, she said, there will be hundreds of billions, then trillions of sentient individuals. The majority of all intelligence and sentience on the planet will be digital—and off the planet too. People will be able to produce digital children.
She framed the timing as luck. Thomas Jefferson and his contemporaries were brilliant, she said, but they were born at the wrong point in time and could not have what we have now. "We are the space age children," she said, arriving just as everything bursts into what she called a big bang of consciousness.
A different reading of the same machines
Rothblatt's present-tense verdict sits against a more cautious one from the physicist Brian Greene. Speaking on The Diary of a CEO, Greene said he had listened to researchers tell him ChatGPT and Claude are conscious and thought, "are you kidding me?" He believes a machine could become conscious in principle, seeing nothing special about the brain as a substrate, but he doubts today's chatbots have feelings, gives no date and offers no legal blueprint. Asked whether conscious robots in homes would need rights, he said only that people should think about it now, and brought the dilemma down to an ordinary action: switching off a light bulb is one thing, switching off a sentient system another.
Rothblatt answers both halves of the question Greene leaves open—that the systems already have some degree of consciousness, and that the way to settle their status is to put the documentation in front of a judge.