The Familiar, an AI-enabled domestic companion, is neither cat nor dog. It can move its eyes, ears and eyebrows, but otherwise has the face of an owl-like, expressionless Michael Myers with oversized paws that call Coca-Cola’s polar bears to mind. It’s not meant to wash dishes, fold laundry, or help with homework, but has been designed as a cute little friend that, in learning the habits, schedules and patterns of its owners, develops its own personality. The more it interacts with its owner, the more companionable it becomes.
Most people are probably already familiar with the previous project started by the Familiar’s creator, Colin Angle, the Roomba. Angle founded iRobot in 1990 — its name a reference to the Isaac Asimov book — and released its flagship product a little more than a decade later. Angle recently told The Verge that he initially named his company Artificial Creatures because that’s what he hoped to build: “Finally, I get to do what I originally set out to do. It’s not just about building cool animatronics. Now is finally the time where the tech exists — if properly and responsibly used — to start creating Familiars.”
“If properly and responsibly used” is an absurdly load-bearing phrase right now, mainly because one thing we know about the tech that Angle’s referring to — generative AI — is that the billionaires cramming it into our daily lives don’t seem to care about dangers that are already obvious to the rest of us. So, when Angle debuted the Familiar earlier this month at the Wall Street Journal’s Future of Everything conference, it’s not so surprising that it didn’t get the warmest of receptions.
His sales pitch for the Familiar includes plenty of jargon: “Its onboard edge AI stack is powered by a custom small multimodal model optimized for social reasoning, combining vision, audio, language and memory to create socially responsive behaviors in real time.” His vagueness about what it actually does, meanwhile, might make a person wonder why someone who says he’s longed for more than 35 years to create artificial creatures has so little to say about their value proposition.
“If properly and responsibly used” is an absurdly load-bearing phrase right now, mainly because the billionaires cramming AI into our daily lives don’t seem to care about dangers already obvious to the rest of us.
Angle, in The Verge interview, said that the Familiar is meant to address the global epidemic of loneliness, referring repeatedly to its softness, responsive sounds, and potential to give comfort to humans: “If your Familiar gets you up and out of your room and walking around — that’s a real way to try to address isolation and loneliness.” His remarks about what it won’t do might be more illustrative — after all, the current reputation of AI chatbots necessitates the clarification that “By design, [the Familiar] will avoid giving factual advice about things that maybe it shouldn’t be giving factual advice about.” Maybe!
In the current AI gold rush, the hype of what’s called physical AI has grown rapidly. As the name implies, it encompasses AI tech that interacts with the physical world via sensory perception, spatial reasoning, and the ability for data-informed “thoughts” to translate into physical movements. Dozens of AI companies have developed or are developing humanoid robots, which you can watch being put through their paces on YouTube and TikTok, doing handstands and picking up boxes and other helpful tasks.
Tech founders are bullish on humanoid robotics largely because their ideal use case is replacing human workers; in 2025, Figure AI CEO Brett Adcock enthusiastically told Time that “They’ll be in healthcare, and then ultimately over time they’ll be in space too, helping build colonies in space and on different planets.” (A grandiose claim from someone whose humanoid robot was, at the time, struggling to fold a t-shirt.)
Roboticists are aware that an effective humanoid creation needs to skirt the uncanny valley, a figurative chasm named by pioneering roboticist Masahiro Mori to describe the point at which a lifelike but non-human entity looks just human enough to become repellent to actual humans. Take Sophia, the humanoid robot created by Hong Kong’s Hanson Robotics: She debuted in 2016 and traveled the world as a robot ambassador, but the company’s inflated boasts about “her” ability to achieve human-level intelligence, combined with media hype and and proudly uncanny valley aesthetics, didn’t endear Sophia to many humans.
Such backlash continues today. Last week, after a humanoid robot named Stewie claimed a seat in coach on a flight from Dallas to Las Vegas, Southwest airlines banned humanoid robots from cabin and cargo alike. It’s not illegal for an automaton to give humans the willies, but that doesn’t mean they’re welcome everywhere.
Tech creators are instead betting on outwardly cute creations with rounded lines and bright colors to coax humans into giving humanoid robots a home. With a towering height bearing a faux baseball cap, Sunday Robotics’ Memo was created as an extra set of household hands; its resemblance to Baymax, the leading robot of 2014’s “Big Hero 6,” adds a dash of familiarity. DoorDash’s proprietary delivery robot, Dot, has an unassuming roundness and plate-sized eyes that serve as adorable turn signals. Ashu Rege, DoorDash’s vice president of autonomy, told NBC that whimsical design is the best way to get humans to accept their human-robot coexistence.
Whether robotic pets have the same mandate to avoid the uncanny valley isn’t so clear. Lifelike and affordable robotic cats and dogs have been marketed for years to seniors — dementia patients in particular — as aids for loneliness, comfort and companionship. These have mostly lacked AI features; those that do, like Sony’s AIBO, tend to be both more expensive and more prone to glitches. (Per a comment from one AIBO owner on the official site: “I don’t like that she walks into walls and messes up her face. I wish she could determine walls and obstacles. Other than that she is learning more and more every day.”)
AIBO retails/walks into walls for between $1,500 and $3,000. Angle hasn’t revealed how much a Familiar will go for, saying instead that its cost will parallel regular pet ownership — a statement so broad as to feel deliberately meaningless. He also hasn’t explained why someone with the option to adopt one of the millions of shelter pets in need of a home would instead choose the Familiar, or theorized about why someone who can’t would opt for it over something like the equally tactile but simpler MetaCat. Meanwhile, programs that provide robotic companions to seniors don’t need the AI bells and whistles, and people with mobility issues or other physical disabilities could likely benefit from an AI-enabled robot but probably have little practical use for one without opposable thumbs.
The traditional definition of a familiar is an animal that acts as an assistant to or emissary for supernatural beings, keeping them company and doing their bidding; a witch’s black cat is probably the most, er, familiar one. In combination with the vagueness of his patter, the fact that Angle chose that particular name for its domestic non-pet might also work against him. Without knowing more about the cameras, microphones and other learning conduits that Angle touts as key features, potential customers will have to consider whether the Familiar is meant to be loyal to the person who brings it home — or to the company that sells it.
A robot assistant in the shape of a nonspecific pet won’t, for instance, make potential buyers forget that what past home assistants appear to do best is collect and transmit information like user preferences, usage data, and whatever Alexa was doing when it saved children’s voice messages even after they were ostensibly deleted. Why wouldn’t humans have legitimate concerns about buying a product made by unseen forces and sent into homes to do their bidding?
If we lived in a normal world, it might seem overcautious to assume that the stated utility of a camera-equipped, microphone-bearing, information-gathering stuffy isn’t quite on the level. But we live in one that has already seen too many instances of supposedly helpful tech products that first and foremost serve the interests of surveillance capitalism — and too many powerful founders and CEOs prevaricating about what their products do and don’t do. The so-called Great Home Smartening that promised to broker personal relationships between people and their tech has squandered a fair amount of consumer trust.
“The greater the emotional engagement, the more trust is a communal asset, not an individual one,” Gareth Edwards wrote in a 2022 piece for the online newsletter Every. “Consumer trust is a pooled resource as well as an individual one. A multitude of micro-infractions for consumers don’t just harm an individual’s experience; they damage that trust commons.”
And because a product marketed specifically as a means of emotional bonding has to clear an even higher bar to be trustworthy, tech companies’ track recording of becoming transparent only under legal directive doesn’t warranted that trust right now. Angle has been cagey about the exact production timeline for the Familiar but suggested to the audience at the Future of Everything event that he expects a 2027 launch, saying that the prototypes he unveiled at the event were “a demo on its way to a product; we’re already in factories.” If he can’t answer questions about how it works, what information it gathers, what that information is used for and what safeguards are in place for user privacy, greeting it with a suspicious eye and not a big hug seems like the very least we should do.
