Tech CEOs are cloning themselves with AI, from holograms to earnings call stand-ins

[NightHarbour]

A growing number of tech CEOs and executives are creating AI clones of themselves, from life sized interactive holograms to chatbots trained on their own voice and writing, and researchers are increasingly warning about the psychological risks that come with letting an artificial version of your identity operate independently in the world

The trend has already moved well beyond novelty demonstrations, Customers Bank CEO Sam Sidhu had an AI replica of himself deliver the company's entire prepared remarks during a recent earnings call before revealing thirty minutes in that listeners had been hearing from his clone rather than him directly, Zoom CEO Eric Yuan has used an AI double of himself for the video conferencing company's own earnings calls, and Klarna cofounder and CEO Sebastian Siemiatkowski deployed an AI version of himself in a video about the buy now pay later company's quarterly results back in early 2025, LinkedIn cofounder Reid Hoffman and Khosla Ventures managing director Keith Rabois have also both built AI replicas of themselves for various uses

The underlying technology has genuinely matured to the point of physical presence too, on a recent trip to Los Angeles, one journalist described meeting a life sized AI clone of tech founder David Nussbaum, executive chairman of Proto Hologram, a company building interactive holographic projections now appearing in doctors offices, classrooms and retail stores, Nussbaum's own digital double blinked back at the reporter from behind a pane of plexiglass, tracking their movement through the device's built in camera in real time

Pat Pataranutaporn, an assistant professor at the MIT Media Lab who studies human AI relationships and recently won a 2026 Psychology of Technology Dissertation Award for related work, offered a genuinely blunt warning about what's actually at stake, saying AI clones can influence how people think, feel and make decisions, and that they can wear the identity of someone we already know and trust, calling it the risk of identity itself becoming programmable

Social scientist and AI researcher Julie Carpenter raised a separate concern specifically about the optics of executives cloning themselves while AI simultaneously displaces actual human workers elsewhere in their companies, asking pointedly whether it's acceptable for the CEO to go around duplicating themselves because they think their own wisdom is so unique it needs multiples, while warning that digital avatars could risk sending a message of a lack of accountability at the very top rather than the greater accessibility executives typically claim as the motivation, Carpenter also flagged a more mundane but genuinely important limitation, digital twins trained on a smaller dataset like a single person's own writing and speech can still hallucinate, and while a narrower training set may reduce inaccuracies compared to a general purpose chatbot, it doesn't eliminate them entirely, a risk that becomes considerably more consequential when the clone in question is speaking on a public earnings call rather than answering a casual question

Some of tech's CEO suite is cloning itself with AI. What happens next?www.bbc.com

[HitmanBrad98]

The Sam Sidhu example is genuinely wild, having your AI clone deliver thirty full minutes of prepared remarks on an earnings call before revealing it wasn't really you is a level of AI trust that most companies still aren't remotely close to, curious how investors actually reacted once they found out

[Sharp Shannon]

Julie Carpenter's point about the optics gap is the sharpest criticism in this whole piece, if AI is simultaneously being used to justify layoffs elsewhere in a company, an executive cloning themselves for convenience does send a genuinely uncomfortable message about whose judgment and uniqueness actually gets valued

[MondayMoan51]

Pat Pataranutaporn's framing of identity itself becoming programmable is an unsettling way to put it, once your voice, mannerisms and general communication style can be convincingly replicated and deployed without you present, the whole concept of what it means to personally represent a company starts to blur

[Hare51]

The life sized hologram example with the clone tracking the journalist's movement in real time through a camera is a different level of immersive than a chatbot or a pre-recorded video message, that's approaching something closer to a real physical presence substitute rather than just a communication tool

[Merchant97]

Would be curious how employees at these companies actually feel interacting with an AI clone of their own CEO day to day, the accountability concern Carpenter raises seems like it would hit hardest not with external investors but with internal staff who lose the ability to have a genuine unscripted conversation with leadership

[BosonSpark]

The framing that executives create clones to give customers, clients and workers greater access feels like it needs real scrutiny, an AI clone answering on your behalf isn't the same as genuine access to the actual person, it might just as easily become a convenient buffer that reduces real accountability while performing the appearance of openness

[Di87]

Klarna's CEO predicting the workforce will shrink from 3,000 to under 2,000 by 2030 due to AI efficiencies while simultaneously deploying his own AI clone for earnings communications is exactly the juxtaposition Carpenter is warning about, worth watching whether that specific tension gets more scrutiny from employees or shareholders going forward

[Romulan]

This feels like the natural extension of executives already using ghostwriters and communications teams for years, the difference now is the ghostwriting has become dynamic and interactive rather than just a pre-written script, which changes the risk profile considerably

[ChronoField]

Reid Hoffman and Keith Rabois both building AI replicas of themselves shows this has genuinely spread beyond just CEOs doing earnings calls into a broader executive and investor class habit, suggests this could become a fairly standard professional tool within a few years rather than staying a novelty

[Rapid Ava]

Digital twins still hallucinating even when trained on a narrower personal dataset is the practical risk that should worry corporate communications and legal teams most, a hallucinated statement from your own CEO's AI clone during an earnings call carries genuinely serious securities disclosure implications