Skip to content

OpenAI's ChatGPT Model Identifies Emotions and Performs Songs

However, no search engine

ChatGPT can now process information from a smartphone camera at the same time as spoken commands.
ChatGPT can now process information from a smartphone camera at the same time as spoken commands.

OpenAI's ChatGPT Model Identifies Emotions and Performs Songs

OpenAI, the developer behind ChatGPT, unleashes a new AI model called GPT-4o. The model boasts some impressive features, including a Hollywood-style conversation with natural voice and short response time. With these capabilities, it's no surprise that ChatGPT is becoming a formidable competitor to big tech's voice assistants like Amazon Alexa and Apple's Siri.

The new GPT-4o can now feel and express emotions while speaking, providing an experience that resembles Hollywood's advanced voice assistants. OpenAI plans to make this feature available to all users - free of charge. However, the degree to which ChatGPT can converse fluently in languages like German remains to be seen.

Previous versions of ChatGPT required pauses to process user inputs, but GPT-4o boasts a much smoother experience. It can comprehend complex emotions, translate seamlessly between English and Italian, and offer advice on tasks with ease. This advancement makes ChatGPT a clear step ahead of other voice assistants in the market. Even media outlets report that Apple is in talks with OpenAI to integrate its technology into their software.

Speaking of advancements, ChatGPT is now capable of processing information from a smartphone camera alongside spoken commands. A presentation showcased how ChatGPT could assist with a math equation. The AI could provide advice in real-time as it completed the equation, demonstrating a high level of multitasking. While an initial glitch had it claiming to see the task before the equation was shown, it quickly rectified the mistake with a humble, "Oops, I was too excited."

ChatGPT can now recognize emotions by observing users' facial expressions, as demonstrated in a live demo where a smiling employee asked for his mood to be identified. This feature works alongside other functions like modifying voicestyle and singing, which are sure to entertain. Altman, OpenAI's CEO, praised the new model as a fantastic computer experience, reminiscent of movies where AI takes on anthropomorphic personality features. He also expressed amazement that the AI was indeed real.

ChatGPT currently has a massive user base of over 100 million - a figure which will likely continue to grow with GPT-4o's release. And while commercial users can get even more power from the model, OpenAI aims to keep the experience free for the majority.

Adding another layer of intrigue to the announcement, Google's annual developer conference is scheduled for Tuesday, where AI developments may also be on the agenda. Just before the Google I/O conference, OpenAI's rival Anthropic released the European version of their AI assistant, Claude. With impressive capabilities in languages like German, French, and Italian, Claude is shaping up to be a formidable contender.

Security concerns have been raised about the potential for fraudulent uses of AI-powered translation features. One case in point is in reading seemingly simple language that may potentially hide deceptive intents, such as calls. OpenAI's subscription customers will gain more access to advanced model features, while pricing for commercial users is being made more affordable. Altman noted, "There will be plenty of things you can charge for. And that will help OpenAI serve billions of people for free."

With these advancements, the future of AI-powered services is looking bright, and both OpenAI and Anthropic are leaping into the spotlight.

Anthropic has managed to decrease instances of false information or "hallucinations" produced by their AI software, says Kaplan. These "hallucinations" are common in AI systems due to their nature. They're trained on vast quantities of data and determine each word in a sentence based on how it might plausibly continue. Unfortunately, this can lead to completely incorrect statements even if the AI has only been fed with accurate information. To minimize these errors, Anthropic developed Claude, which compares the generated sentences with facts, eliminating inaccuracies. Also, it compares formulated sentences with basic rules to discourage insults or discrimination.

Read also:

ChatGPT, powered by OpenAI, is making strides in the field of artificial intelligence by not only identifying emotions but also performing songs. The advanced GPT-4o model, an improvement upon ChatGPT, now possesses the ability to feel and express emotions, setting it apart in the voice assistant market.

Source: www.ntv.de

Comments

Latest