OpenAI is reportedly taking away ChatGPT-4o’s Scarlett Johansson-like voice that it says is not meant to be an imitation of the actress.

The company will pause the ‘Sky’ voice in its latest version of ChatGPT after people said it sounded like Johansson, who voiced an AI companion in the film Her, Bloomberg reported.

“We believe that AI voices should not deliberately mimic a celebrity’s distinctive voice — Sky’s voice is not an imitation of Scarlett Johansson but belongs to a different professional actress using her own natural speaking voice,” OpenAI said, adding that it would not share the names of the voice actors it worked with to protect their privacy.

ChatGPT-4o, which was released last Monday, has five distinct voices that were “carefully selected through an extensive process spanning five months” from over 400 submissions, OpenAI said in a blog post on Sunday. The company said it started working with “professional voice actors, talent agencies, casting directors, and industry advisors” in early 2023 to choose the voices: Breeze, Cove, Ember, Juniper, and Sky. Some of the criteria for ChatGPT’s voices included speaking multiple languages and “natural and easy to listen to” voices.

To audition, OpenAI said actors were given scripts of ChatGPT responses, ranging from “answering questions about mindfulness to brainstorming travel plans, and even engaging in conversations about a user’s day.”

“We spoke with each actor about the vision for human-AI voice interactions and OpenAI, and discussed the technology’s capabilities, limitations, and the risks involved, as well as the safeguards we have implemented,” OpenAI said. “It was important to us that each actor understood the scope and intentions of Voice Mode before committing to the project.”

OpenAI said it will continue to compensate each voice actor at “above top-of-market rates” as long as their voices are used in the company’s products. The company did not immediately respond to a request for comment.

The company said it plans to give ChatGPT Plus users access to a new Voice Mode for GPT-4o in alpha in the coming weeks.

This article originally appeared on Quartz.

Shares:

Leave a Reply

Your email address will not be published. Required fields are marked *