OpenAI Halts Scarlett Johansson Voice After Her Objection

Image credit: Photograph: Sarah Meyssonnier/Reuters

Scarlett Johansson has criticized OpenAI for using a voice in its new ChatGPT product that closely resembles hers without her consent. Johansson revealed that OpenAI approached her nine months ago to voice the AI system, but she declined for personal reasons. She expressed shock and anger when she discovered the voice option, called “Sky,” sounded so similar to hers that friends and news outlets couldn’t distinguish the difference.

Following the backlash, OpenAI removed Sky from ChatGPT. The voice had been showcased during the launch of ChatGPT-4o, with demonstrations highlighting Sky’s personable and responsive nature, drawing immediate comparisons to Johansson’s AI character in the 2013 film “Her.” OpenAI CEO Sam Altman hinted at this similarity with a tweet saying “her.”

OpenAI clarified that Sky’s voice was not based on Johansson but belonged to a different professional actress. The company emphasized its respect for the voice acting industry, stating Sky’s voice actor used her natural speaking voice and remained anonymous to protect her privacy.

Johansson recounted that Altman initially approached her, believing her voice could bridge the gap between tech companies and creatives, making AI more comforting to consumers. He reached out again two days before Sky’s release, prompting Johansson’s lawyers to demand the voice’s removal.

The similarity to Johansson’s character in “Her” sparked commentary, including a segment by her husband, Colin Jost, on “Saturday Night Live.” Critics also questioned the voice’s overly feminine and fawning nature, with “The Daily Show” host Desi Lydic joking about the gendered design.

OpenAI defended its selection process, explaining that ChatGPT’s voice options were chosen for their timeless quality and trustworthiness. The company reviewed hundreds of voice acting submissions and released five different voices for ChatGPT in September. The chosen actors recorded their voices in San Francisco to train the models.

The removal of Sky’s voice came shortly after several top safety team members resigned, including key researcher Jan Leike, who accused the company of prioritizing flashy products over safety. Altman and co-founder Greg Brockman maintained that safety concerns would prevent product releases.

Open AI’s blogpost highlighted its collaboration with entertainment professionals and compensation for voice actors. The company, like others in AI, faces intense scrutiny and legal challenges from entertainers, creators, and media companies over copyright issues and concerns about AI replacing human jobs. Major entertainment unions, such as SAG-AFTRA, have struck over AI usage and likeness rights.

Re-reported from the article originally published in She the People.

Leave a Reply