The application, filed with the U.S. Patent and Trademark Office on Tuesday, came a day after Altman's interview discussing OpenAI's plans to unveil multiple significant releases preceding its highly anticipated GPT-5 model. While trademark filings don't guarantee a product launch, OpenAI is widely expected to deliver an enhanced version of its viral ChatGPT software in mid-2023.
In a conversation with podcaster Lex Fridman, Altman teased an "amazing new model" for 2023, stating OpenAI will roll out "many different things" in the coming months, promising exciting developments. He hinted at releases distinct from GPT-5, suggesting a strategic roadmap beyond OpenAI's flagship language model.
Currently, OpenAI lacks a consumer-facing digital voice assistant, though it offers speech recognition capabilities through APIs like TTS for speech-to-text conversion and Whisper for general speech recognition.
Alongside the voice assistant filing, OpenAI secured trademarks last October for GPT-6 and GPT-7, covering expansive use cases such as simulating conversations, predictive analytics, music generation, and code creation via artificial intelligence.
As OpenAI charts an ambitious path, it navigates regulatory hurdles; the Patent Office recently rejected the company's attempt to trademark "GPT," deeming the acronym too descriptive.
The "voice engine" application specifies ten intended use cases, including building voice assistants, speech recognition, audio generation from text prompts, multilingual transcription, and voice service development underpinned by machine learning.
Read next: Authoritas Analysis Warns of Organic Traffic Decline Amid Google's AI-Based Search Replies
by Asim BN via Digital Information World
No comments:
Post a Comment