- Reid Hoffman gave an interview with a deepfake of himself.
- While the co-founders of LinkedIn acknowledged the dangers of twin AI, they also explored its benefits.
- The bot discussed AI regulation and spoke in Klingon during the conversation.
LinkedIn co-founder Reid Hoffman's interview with an artificially intelligent version of himself started with a Jerry Seinfeld impression and a conversation in Klingon, and it only got creepier as the conversation progressed.
The investor, who is a vocal supporter of AI, said he wants to experiment with the technology and videos while learning how it is challenged by deepfakes, known as lead AI. Despite his pro-AI stance, he admitted he was hesitant about the interview.
“At first I thought I wouldn't like this,” Hoffman said. “But I actually thought it was kind of interesting.”
The bot was built on OpenAI's GPT-4 and trained on more than 20 years' worth of material from Hoffman's speaking engagements and published books. Reed AI and Hoffman asked each other questions in an interview published Wednesday.
During the conversation, the two leads discussed AI regulation, its capabilities, and how Hoffman can improve his LinkedIn profile. When Hoffman was asked about the ethics of deepfakes, he said setting “rules of the road” is important for both public and private citizens.
Lead AI also asked about the risk of technology taking jobs away from real people, ironically. The real Hoffman likened it to the invention of the steam engine.
“Even if there is some pain in the transition, we win by being the first to accept it, learn early and implement it as a society,” he said.
Surprisingly, Lead AI has shown support for governments regulating the technology itself, even as advances continue.
“We need a framework that not only fosters innovation, but also ensures that the benefits of AI are equitably shared, with a focus on improving the public good,” Reed AI said.
As technology advances, the billionaire has become a champion of deploying AI to “improve humanity.” Hoffman told Time magazine that “blitzscaling,” a strategy that prioritizes business growth above all else, should be applied to AI.
