close
close

OpenAI delays Sky, its new AI voice, for security reasons

OpenAI delays Sky, its new AI voice, for security reasons

  • OpenAI announced that the launch of a new voice for ChatGPT called Sky has been postponed due to security concerns.
  • The controversy arose due to the fact that she was very reminiscent of Scarlett Johansson’s voice in the movie “Her,” in which she played an AI assistant.
  • OpenAI only said that the voice would be trained to reject certain requests if they violate company policies.

OpenAI delays Sky, its new AI voice, for security reasons

OpenAI announced that the release of a voice called Sky, which caused controversy last month, has been postponed and OpenAI fans are not happy about it.

Last month, OpenAI released a set of new voices for ChatGPT at an event, and one of them sounded dangerously similar to Scarlett Johansson’s voice in the movie “Her,” in which she played an AI assistant.

Apart from the similarity in voice, it is flirtatious nature also led to a lot of controversy. Things got so bad that at one point the company decided to remove the voice.

However, it seems that OpenAI has overcome these concerns and still plan to release itbut with a delay due to some security concerns. Apparently the voice still needs to go through a few tests.

What are the security concerns?

OpenAI hasn’t revealed much about its plans for voice, but it did mention Programming to reject certain requestsIf that is the case, it is nothing new.

Since “Sky” was a bit flirtatious in the demo, there is a higher risk of abuse.

Certain users with malicious intent are known to ask controversial questions to AI chatbots or make them say something offensive. Even if it is the user’s fault, at the end of the day, ChatGPT will be criticized online when something like this happens.

It is also reported that the The infrastructure to support the tool is not yet ready to handle millions of users and their requests.

“As part of our iterative deployment strategy, we will launch the alpha version with a small group of users to collect feedback and expand it based on our learnings” – OpenAI

The exact release date has yet to be announced, but it is expected to launch in July Testing the voice with the small group and every user is expected to have access to it by fall.

A stunt gone wrong

On the other hand, it is quite possible that this “update” is just a Excuse to keep Sky away from the starting blocks for some time. Scarlett Johansson was obviously not happy with OpenAI afterward and threatened to take legal action against them. And she has good evidence to explain the connections.

Sam Altman contacted Scarlett about 9 months ago to Ask them for permission to use their voice for an AI assistant. However, she declined due to personal reasons. Even a day before the launch, the OpenAI team tried to reach Scarlett but received no response.

Even after rejection Sam started Sky anyway. But people quickly noticed the uncanny resemblance between Sky and Scarlett. OpenAI tried to dismiss it as a coincidence and assured that there was no foul play. But it’s not hard to see through what’s going on here.

OpenAI tried to cover up the matter, publishing a lengthy blog explaining how the voice actors were chosen. Sam Altman said they chose the voice actor for Sky before approaching Ms Johansson.

This just goes to show that Sam was very aware of what he was doing. He knew that the voice was similar to Scarlett’s and tried to get her to work with him officially. But it didn’t work out.

OpenAI has since removed Sky from ChatGPT “out of respect for Ms. Johansson.” However, if they wanted to “respect” her, the voice would not have been published in the first place.

OpenAI user reaction

However, OpenAI fans who had been eagerly awaiting the new voice are not happy with the news. Some of them have even feel cheated because they paid for the ChatGPT Plus subscription believing that the new voice will be available within a few weeks. Others mock the company for not keeping its promise and not delivering the products on time.

The timing of this news is particularly unfortunate because it Rival Anthrop has just released a new update for its chatbot Claude, now Claude 3.5.

This does not paint a good picture for OpenAI. The competition in the industry is far too fierce to make such mistakes. One wrong step and other competitors will crush it beneath their feet and pull away.

The Tech Report - Editorial ProcessThe Tech Report - Editorial ProcessOur editorial process

Tech Report’s editorial policies focus on providing helpful, accurate content that provides real value to our readers. We only work with experienced writers who have specific knowledge of the topics they cover, including the latest developments in technology, online privacy, cryptocurrencies, software, and more. Our editorial policies ensure that every topic is researched and curated by our in-house editors. We maintain strict journalistic standards, and every article is 100% written by real writers.