OpenAI Takes Action Against Unauthorized Deepfakes
OpenAI, under the leadership of Sam Altman, is stepping up its efforts to tackle unauthorized deepfake content generated by its Sora 2 text-to-video tool. This move comes in response to concerns raised by various public figures, including Bryan Cranston of “Breaking Bad” fame.
Since the launch of Sora 2 on September 30, numerous unlicensed deepfake videos have surfaced on social media, featuring celebrity voices and images without proper attribution or compensation. This has understandably led to complaints.
Cranston himself was featured in a fake video where he appears to converse with Michael Jackson. He brought the issue to the notice of SAG-AFTRA, which has prompted OpenAI to take action, as announced by the actors’ guild on Monday.
“We appreciate OpenAI’s enhanced policies and safeguards and hope that they, along with other companies, will honor our rights to control how our voices and images are used,” Cranston remarked in a joint statement with the union and OpenAI.
Recently, OpenAI blocked users from generating deepfakes of Martin Luther King Jr. after his estate criticized the portrayal as “disrespectful.” This has also raised further awareness regarding the ethical use of AI-generated content.
Zelda Williams, daughter of the late actor Robin Williams, previously urged the public to refrain from creating deepfakes of her father using Sora. This highlights the personal impact of such technology.
OpenAI has indicated that Sora will enforce a stricter “opt-in” policy, necessitating explicit permission from celebrities before utilizing their likenesses or voices in AI-generated videos.
In a statement released on Monday, the company pledged to “quickly respond to any complaints” regarding potential misuse in the future.
Hollywood agencies CAA and UTA, which had expressed concerns about potential risks to their clients and intellectual property due to Sora’s functionality, have also collaborated with OpenAI to ensure greater accountability.
Altman reaffirmed the company’s support for the NO FAKES Act, a federal initiative aimed at prohibiting AI-generated videos that depict individuals without their consent.
“OpenAI is dedicated to safeguarding our performers against the misuse of their likenesses and voices,” Altman stated. “We were among the early advocates for the NO FAKES Act, introduced last year, and we will continually uphold the rights of performers.”
