Leonardo Ai pledges to strengthen guardrails after deepfake porn investigation
The Sydney-based generative AI company raised $47 million in December and has 7 million users.
Sydney-based generative artificial intelligence startup Leonardo Ai has pledged to ramp up efforts to tackle deepfake porn created on its platform after revelations some users were bypassing restrictions on creating nonconsensual sexual content.
Leonardo users in a Telegram group dedicated to deepfake porn shared tips on how to use the company's text-to-image generator to create sexual images of celebrities like Billie Eilish, US-based 404 Media found.
"We are continuing to add further logic and mechanisms to better address methods that have been used to sidestep [our] guardrails," a Leonardo Ai spokesperson said. "Many of these changes are already live and more are coming over the next couple of weeks."
Leonardo Ai was founded in 2022 as a tool to create character models in video games, but pivoted to become a competitor to Midjourney and OpenAI's Dall-E. The company hopes to beat out rivals with features like “realtime canvas” mode, which converts Microsoft Paint-esque squiggles into art, and a “Motion” capability that can turn static art into a moving short video.
