This is a screenshot of the live video of the US Open Artificial Intelligence Research Center (openai) releasing gpt-4o
The US Open Artificial Intelligence Research Center (openai) demonstrated the latest version of chatgpt on the 13th: start voice conversations with users, recognize images and start discussions, Translation... Compared with the previous version, it basically has no delay in talking to users. It can listen, chat, and is good at observing words and expressions. People exclaim that the new version of chatgpt is "more human". What breakthroughs has the new version of
chatgpt made? In what fields can it be applied? How big an impact will it have on the field of artificial intelligence? The reporter will help you solve the problem - what are the "evolutions" of
? The artificial intelligence model released by the
Open Artificial Intelligence Research Center on the 13th is called gpt-4o. The letter "o" in the number represents "omni" (omnipotent). It can accept any combination of text, audio and image input, and can also generate Relevant responses to text, audio and images. The center
not only conducted live demonstrations around gpt-4o, but also released more videos "showing off skills" on social media.
In the video above, gpt-4o sounds obviously "more chatty" and even throws in a few jokes from time to time. There is a slight change in its tone, and there is a smile in its words, making chatting with it more like talking to a real person.
Live chat is a key skill of chatgpt. Compared with previous versions, the main differences of gpt-4o are: first, users can interrupt the chatbot at any time without waiting for it to finish speaking; secondly, it will respond to questions in real time, and there is no longer a need to wait for it to finish speaking. There is a time lag of three seconds; thirdly, it can sense people's emotions. For example, if the presenter is breathing rapidly, it will ask the other person if they need to stabilize their emotions a little. In addition, gpt-4o can generate different styles of sounds. In the
demonstration, gpt-4o used its visual and speech capabilities to guide the demonstrator to solve an equation problem on paper step by step, rather than giving the answer directly. It also demonstrated its ability to translate between English and Italian and recognize emotions from selfie photos. When a presenter told it that he was showing "how useful and incredible" it was, it responded: "Oh, stop it, I'm so embarrassed." Sam, CEO of
Open Artificial Intelligence Research Center Altman wrote in his blog that day that gpt-4o is like the artificial intelligence technology in the movie. "Talking to a computer has always felt unnatural to me, but now it feels natural." What is the application potential of
?
Liu Wei, director of the Human-Computer Interaction and Cognitive Engineering Laboratory of Beijing University of Posts and Telecommunications, said that gpt-4o has shown good capabilities in tasks such as text generation, question answering, question and answer systems, or sentiment analysis. This technological breakthrough will undoubtedly have a major impact on related companies at home and abroad. It not only promotes the development of natural language processing technology, but also makes the application of artificial intelligence in many fields more extensive and in-depth.
He believes that gpt-4o will promote the application of artificial intelligence in financial services, education, medical care, driverless cars and other fields, and is expected to lead the comprehensive development of artificial intelligence technology.
In recent years, competition to develop more user-friendly and powerful generative artificial intelligence tools has been fierce. Just the day after the Open Artificial Intelligence Research Center released gpt-4o, Alphabet Inc., Google’s parent company, held an annual Google developer meeting, with artificial intelligence being the highlight. Silicon Valley entrepreneur Elon Musk and Mustafa Suleiman, one of the founders of the technology company DeepThink, have also invested in the development of chatbots grok and pi respectively, with anthropomorphic features as the main product focus.
BBC comments that gpt-4o is able to combine text, audio and image content to react instantly and is currently leading the competition. Mira Murati, chief technology officer of the
Open Artificial Intelligence Research Center, said that gpt-4o is scheduled to go online within a few weeks, and users can try it for free. Original paying users of chatgpt will receive more permissions to use the new version. What should
worry about?
In a demonstration video released by the Open Artificial Intelligence Research Center, gpt-4o guides a boy step by step to solve math problems.Some netizens said that they no longer have to help their children with homework. Some people are worried about whether the teaching profession will be replaced by artificial intelligence.
Some artificial intelligence experts believe that while the new version of chatgpt may be more advanced and easier to use than its competitors, it is unlikely to completely eliminate some professions anytime soon. People who work in teaching or translating are more likely to use these tools rather than replaced by them. Leslie Theo, senior director of artificial intelligence products at
Singapore Artificial Intelligence Initiative, said that teaching involves human empathy. "Teachers themselves have gone through the learning process and understand how people overcome difficulties, but artificial intelligence is different." He believes that jobs such as teaching, translation and customer service are unlikely to disappear due to the emergence of gpt-4o.
There are also some experts who believe that technical demonstrations can cause reactions in most people. They are usually carefully crafted and may not reflect the true functionality of the product.
Liu Wei said that the voice function of gpt-4o has changed the game rules of conversational artificial intelligence, but it still has not achieved the deep situational awareness capabilities of chat robots, such as intention understanding and motivation analysis. In addition, the progress of gpt-4o has brought new challenges in the ethics and security of artificial intelligence, requiring us to carry out new thinking in terms of data privacy, information credibility, potential bias, malicious use, awareness and responsibility.
Source丨Xinhua International Headlines (Copyright belongs to the original author, if there is any infringement, please contact us to delete)
