Public Citizen Inc.

05/21/2024 | Press release | Distributed by Public on 05/21/2024 09:24

OpenAI’s Dust-Up Over Scarlett Johansson’s Voice Proves Urgent Need for AI Regulations

May 21, 2024

OpenAI's Dust-Up Over Scarlett Johansson's Voice Proves Urgent Need for AI Regulations

AI corporations will not restrain themselves. Congress must act.

WASHINGTON, D.C. - Earlier this month, OpenAI announced the rollout of its Voice Mode feature for ChatGPT. Many people noted that one of the optional voices sounded like the actress Scarlett Johansson, who had embodied an AI voice in the movie Her. Johansson revealed yesterday that OpenAI had sought to use her voice for Voice Mode, but she had declined. In a statement, she said she was "shocked, angered and in disbelief" about the voice that resembled hers.OpenAI has statedthat it did not use Johansson's voice and that it ispausing the useof the voice that resembles Johansson's.

In response, Public Citizen president Robert Weissman said:

"OpenAI owes Scarlett Johansson more than ahalf-hearted apology('We are sorry to Ms. Johansson that we didn't communicate better').

"It also needs to make amends to people in the United States around the world.

"That should mean:

  1. Pausing its product development and rollouts to grapple seriously with how it respects and compensates creators, how it ensures it does not ripoff people's names and likenesses, and how it avoids appropriating the internet commons.
  2. Suspending its Voice Mode ChatGPT feature. Voice Mode is an overt attempt to anthropomorphize artificial intelligence - a gigantic and enormously risky social experiment given theknown hazards and perils. AsGoogle researchersnote, these include over-reliance on and undue trust in the AI, emotional manipulation, loss of privacy, and undermining of human-human relationships.

"This incident should be a wake-up call to Congress. The purportedly 'responsible,' purportedly non-profit AI firm is rushing recklessly ahead with product rollouts that are insufficiently vetted and insufficiently tested. The lesson is plain: AI corporations will not restrain themselves. It is up to Congress to impose restraints to ensure AI benefits the public, not a handful of corporations at the expense of the public."