-1715497793794_m.webp)
Soon after rumours around an OpenAI search engine, the latest report from The Information hints the AI giant is preparing a technology that can talk to users via sound and text. It may get object and image recognition features. Plus, it could offer better capabilities across logical reasoning and more compared to the current offerings.
Citing sources, the report noted the company showcased the capabilities to some users. The virtual assistant could take on Apple Siri and Google Assistant. This product could be showcased on Monday at the 'Spring Updates' event. The model may have faster and more responsive image and audio interpretation than the current transcription and text-to-speech models.
Citing The Information, The Verge reported it may help customer executives understand the "intonation of callers" about if they are acting sarcastic or help students with mathematics and translate real-world signs. The product could outperform GPT-4 Turbo in some responses but may deliver wrong answers confidently.
As spotted by Developer Ananay Arora, OpenAI may bring calling capabilities to the digital assistant. However, readers should note that this would likely still not be a part of GPT-5 if announced on May 13. However, The Information reported that the updated underlying technology could be announced by the end of 2024.
Meanwhile, OpenAI CEO Sam Altman recently denied reports of a possible launch of a search engine at the Monday event. The company highlights it will showcase "ChatGPT and GPT-4 updates" at 10:00 AM PT. These updates come one day before the much-awaited Google I/O 2024 where the tech giant could unveil a host of AI features.
On a related note, OpenAI recently unveiled Model Spec, which details how the company's models work. Under the 'Don't respond with NSFW content', the company noted: "We're exploring whether we can responsibly provide the ability to generate NSFW content in age-appropriate contexts through the API and ChatGPT." You can read more on this here.