Generation AI voice and video platforms are fighting a flood of user-generated content

In conversation with Assets lead creator Jeremy Kahn Fortune’s At the Brainstorm AI conference in London, Synthesia CEO Victor Riparbelli and ElevenLabs CEO Mati Staniszewski said they’re still excited about make sure that the voice cloning and video generation technologies their corporations provide be used for a very good purpose.

“As with most new technologies, we immediately address what can go wrong. In this case that’s true – that’s being exploited by bad people to do bad things, that’s for sure,” said Riparbelli, co-founder of the video generation platform.

The use of AI-generated audio or video to mimic public figures has change into a controversial topic. In recent months there have been explicit deepfakes from Taylor Swift and faux robocalls from Joe Biden, most notably observers worry about how AI-generated content will influence the US presidential election this fall.

OpenAI recently announced that it could delay the launch of its AI voice cloning tool on account of risks of misuse, citing the potential political implications: “We recognize that generating speech that resembles human voices poses serious risks which might be particularly distinguished in an election 12 months. “, it said in a press release.

Staniszewski, co-founder of AI voice cloning company ElevenLabs, said Fortune’s Kahn said his company invests in “Know Your Customer” protocols and mandatory disclosures to make sure that all content generated on the platform is tied to a particular user’s profile. The company can be exploring other ways to make it clear to users which content is AI-generated and which is just not, he said.

“All content generated by ElevenLabs can be traced back to a specific account,” Staniszewski said. “One thing we’re committed to is going beyond watermarking AI and real content.”

Riparbelli said Synthesia has protocols in place that require users to confirm their identity and supply consent before creating AI-generated videos.

“Nowadays it is impossible to record a YouTube video and make clones of someone [on Synthesia]. This is how we take control,” Riparbelli said. “We have pretty extensive content moderation and rules about what you can and can’t create.”

A reporter asked in regards to the potential risks of audio deepfakes regarding London Mayor Sadiq Khan. He was the goal of a viral audio clip wherein he impersonated himself and criticized pro-Palestinian marches last November.

“Parliament needs to wake up and realize that by doing nothing it is giving troublemakers the opportunity to be bolder.” Khan told this BBC.

“All content on the market ought to be often known as AI-generated, and there ought to be tools that allow you as a user to quickly get to that information… in order that Sadiq Khan can send a message and we will confirm that it’s an actual message “,” said Staniszewski.

Riparbelli said it can likely take time for the industry and lawmakers to succeed in a consensus on best use and regulate tools like those offered by his and Staniszewski’s corporations.

“As with any new technology, there will be those years where people figure out what is right and what is wrong,” Riparbelli said.