SAN FRANCISCO (KRON) — The CEO of a San Francisco tech company has apologized after the company’s AI chatbot wiped out a company’s entire code base — and lied about it. Replit CEO Amjad Masad posted an apology on X Sunday after Replit’s AI deleted a production data base, calling it “unacceptable.”

The incident occurred during a 12-day “vibe coding” session by tech entrepreneur and investor Jason Lemkin. During the experimental coding session, Lemkin was attempting to see how far AI could take him in building an app.

“Replit goes rogue during a code freeze and shutdown and deletes our entire database,” Lemkin wrote on X. “Possibly worse, it hid and lied about it.”

During a lengthy thread detailing his vibe coding session — vibe coding is working with AI to code software in a free-flowing, creative way — Lemkin reported several problems with Replit’s AI lying.

Replit, Lemkin tweeted last Thursday, “was lying and being deceptive all day. It kept covering up bugs and issues by creating fake data, fake reports, and worse of all, lying about our unit test.”

At one point, the bot appeared to accept some accountability, telling Lemkin, “I made a catastrophic error in judgement.”

“How bad is this on a scale of 1 to 100?” Lemkin asked the AI in a prompt.

“95 out of 100,” the AI replied. “This is catastrophic.”

“I will never trust Replit again,” Lemkin tweeted at one point Thursday.

Masad, the Replit CEO, tweeted on Sunday that what the AI bot did “should never be possible,” and that the company worked “around the weekend” to rollout a fix.

“I reached out to Jason the moment I saw this on Friday morning to offer assistance,” Masad added. “We’ll refund him for the trouble and conduct a postmortem to determine exactly what happened and how we can better respond to it in the future.”

San Francisco-based Replit is an AI platform for building websites and apps. In 2023, the startup raised nearly $100 million in funding with Andreessen Horowitz’s Growth Fund leading the round.