OpenAI, OpenAI hack

In the great generative-AI race, the major players seem to leapfrog one another almost daily with faster models and features but sometimes the need for speed has an adverse impact on quality.

The need for more safety testing has delayed this month’s launch of OpenAI’s latest voice assistant for ChatGPT, called Voice Mode, the company said in a statement on X. A broad release of the assistant is now due in the fall.

“ChatGPT’s advanced Voice Mode can understand and respond with emotions and nonverbal cues, moving us closer to real-time, natural conversations with AI,” OpenAI said in the statement. “Our mission is to bring these new experiences to you thoughtfully.”

OpenAI said it needs more time to make a new voice version of its chatbot better at noticing and blocking some content. The company is also working on rolling out new video and screen-sharing capabilities.

“We’re improving the model’s ability to detect and refuse certain content,” OpenAI said. “We’re also working on improving the user experience and preparing our infrastructure to scale to millions while maintaining real-time responses.”

The delay presents a potential setback for OpenAI as it tries to maintain pace in an increasingly bruising race among AI rivals like Anthropic and Google. When OpenAI presented the new feature in a demo in May, it promised a faster chatbot with more powerful image-recognition capabilities. 

The May demo sparked enthusiasm from ChatGPT users, but prompted the threat of legal action from actress Scarlett Johansson. She claims OpenAI ripped off her voice for one of its AI personas.

That may be the least of OpenAI’s concerns.

With Anthropic and Google pumping out new features — Anthropic just announced new enterprise features for a much-faster Claude 3.5 Sonnet — OpenAI and others are feeling the heat to roll out new features for users.