As the production and use of AI products has boomed in the last year, considerations about how to safely use and develop AI products have become a top topic of conversation and concern. In December, Stanford researchers reported that they’d identified over 1,000 child exploitation images in a widely used image database used to train AI models, including Stable Diffusion 1.5. And heading into the 2024 election, the potential for disinformation stemming from AI technology has weighed heavily across the U.S.

Emad Mostaque, Stability AI’s founder and CEO, said the company wants “to spread good practices and defend society and democracy against misuse of this technology.”

In an interview with NBC News, Irwin said that when she first joined Stability AI she was impressed by the integrity work that was already occurring, like developing filters around datasets. She added that she was also excited by initiatives at the company around disinformation, including research and development into identifying and labeling synthetic content. 

“We’re in the early stages of a big technological revolution,” she said. “And so it’s just awesome to be a part of that.”

In October, Irwin told NBC News that working at X was the hardest experience of her career. She resigned after Musk criticized moderation actions taken under her leadership pertaining to misgendering. An anti-trans film was moderated on the platform, which Musk reversed.

Irwin said she left the company when it became clear to her that there was “there was no longer alignment” between the company and her “nonnegotiable principles.”