They'll be coming for the CEO's soon to save money.....

Workers may fear losing their jobs to AI, but CEOs and other executives are using the tech to "scale" themselves.
Why it matters: Digital doubles can boost productivity and influence, but they can also hallucinate and supercharge inauthenticity.
How it works: Execs use startups like Delphi and Tavus to upload their writings, keynote speeches, interviews and even their meetings to create text and voice chatbots and video avatars trained on their work.

Executives love to clone themselves
Workers may fear losing their jobs to AI, but CEOs and other executives are using the tech to "scale" themselves.
Why it matters: Digital doubles can boost productivity and influence, but they can also hallucinate and supercharge inauthenticity.
How it works: Execs use startups like Delphi and Tavus to upload their writings, keynote speeches, interviews and even their meetings to create text and voice chatbots and video avatars trained on their work.
- Otter CEO Sam Liang says he's so inundated with meetings that he ends each day too drained to talk. "I'm double-booked. Triple-booked." So he asked: Could Otter create a Sam Bot to attend the meetings for him?
- "I prefer to call it Sam's Avatar," the human Liang told Axios. "Avatar feels more human."
- Otter's tool, dubbed "Meeting Avatar" is currently only a prototype, but eventually it will be able to answer questions and engage in meeting discussions. It will also record meetings and send notes and summaries.
- Some CEOs are using clones internally. Others make them public. Execs put their digital doppelgangers on investor calls, use them to chat with new employees about company culture, and send them to important video meetings so they don't have to attend themselves.
- Dara Ladjevardian, co-founder and CEO of AI cloning platform Delphi, puts a link to his "digital mind" chatbot in his email signature. I chatted with his clone before getting on the phone with the real thing.
- It was helpful, but not 100% accurate, which is a rough definition of most genAI right now. I asked the bot if Ladjevardian had access to our chat and it said no, it was totally private. But when I asked the human Ladjevardian the same thing, he said he could access the chats. He explained that it was probably a "bug."
- C-suite engagement with AI is absolutely critical, John Levitt, co-founder and COO of elvex, an enterprise AI orchestration platform, told Axios. "I can absolutely see a difference between customers that have strong engagement and usage from exec leaders versus those that don't."
- The willingness to trust generative AI as a stand-in for their body of knowledge shows that the C-suite believes in the tools they're requiring employees to use.
- "It's sort of widening that gap between executive privilege, technical privilege and the precarity of laborers," says social scientist and AI researcher Julie Carpenter.
- AI might be displacing actual human workers, "but it's OK for the CEO to go around duplicating themselves because they think they're so special, their wisdom is so unique that they need multiples of themselves?"
- "It could risk sending messages either to the public or to their employees that there is a lack of accountability at the very top," Carpenter told Axios.
- Digital twins can also hallucinate. Training a chatbot on a smaller set of data, like a single person or company's writing, videos, and podcasts, could mean fewer inaccuracies, but it doesn't mean none.
- "I find it very fascinating," Box CEO Aaron Levie says. "But to be clear, I probably wouldn't do it for myself."
- Instead, Box is working on digital agents that can contain the entire knowledge base of a company to create "a digital memory of infinite scale."
- Levie believes in AI's ability to "blast past the bottlenecks" across organizations, rather than simply giving people more access to the CEO's mind.