Openai claims new GPT-5 model promotes Chatgpt to “doctoral degree”

Chatgpt-Maker Openai unveils the latest long-awaited version of the artificial intelligence (AI) chatbot GPT-5, which says it can provide doctoral-level expertise.

Sam Altman, co-founder and CEO of Openai, has been called “smarter, faster, and more useful”, saying the company’s new model is in a new era.

“I think having something like the GPT-5 is almost unimaginable at any time in human history,” he said before the release on Thursday.

The issuance of GPT-5’s “doctoral degree” capabilities in areas such as coding and writing and claims to acquire state-of-the-art AI Chatbot as technology companies continue to compete.

Elon Musk recently made a similar claim about his AI chatbot Grok, which has been plugged into X (formerly Twitter).

During the launch of Grok’s latest iteration last month, Musk said it “better than a doctorate in all respects” and called it the world’s “smart AI.”

Meanwhile, Altman said Openai’s new model will suffer less hallucination – a phenomenon in which large language models form the answer and are less deceptive.

Following the trend among major AI developers in the United States, Openai also puts GPT-5 into coders, including Anthropic, whose Claude code targets the same market.

OpenAI emphasizes the overall creation software of GPT-5 and demonstrates the ability to better reasoning capabilities – the answers show that work, logic and reasoning.

The company claims it has received more honest training to provide users with more accurate responses, and says that overall, it feels more human.

According to Altman, the model is “much better”.

“GPT-3 is like talking to a high school student to me…4 feels like you’re talking to a college student,” he said in a briefing before Thursday.

“The GPT-5 is the first time that it really feels like talking to an expert on any topic, like a PhD expert.”

But for Professor Carissa Véliz of the Institute of Ethics of AI, the release of the GPT-5 may not be as important as its marketing suggests.

“These systems are impressive and cannot be truly profitable,” she said. They also point out that they can only imitate – rather than truly imitate human reasoning abilities.

“People are worried that we need to keep the hype or the bubble may burst, so it may be mainly marketing.”

A moral expert said the launch of GPT-5 strengthens the gap between AI capabilities and our ability to manage it in a way that the public expects.

“As these models become more capable, the need for full regulation has become more urgent,” said Gaia Marcus, director of the Ada Lovelace Institute.

BBC AI correspondent Marc Cieslak received exclusive access to the GPT-5 before it was officially launched.

“Experience is similar to using an old chatbot: give it a task or ask questions by typing text prompts.

Now, it is powered by what is called a reasoning model, which actually means it is more difficult to solve problems, but it seems more like an evolution than a revolution. ”

The launch of GPT-5 also has an impact on commercial enterprises that focus on the use of its content.

“As AI content becomes more persuasive, we need to ask ourselves – do we protect the people and creativity of the people we see every day?” said Grant Farhall, chief product officer at Getty Images. “Authenticity matters – but it’s not free.”

Farhall said it is important to double-check how AI models are trained and to make sure they are compensated when using the work.

The company will launch models to all users starting Thursday.

In the next few days, is it really as good as Sam Altman said.

Anthropic recently revoked Openai’s access to its application programming interface (API), Require The company violated its Terms of Service by using its coding tools before the launch of GPT-5.

An OpenAI spokesman said evaluating other AI systems to assess their own progress and security is the “industry standard.”

They added: “While we respect Anthropic’s decision to cut off API access, it’s disappointing considering that our API is still usable.”

With the free hierarchy of its new model, the company may indicate a potential shift from proprietary models that previously dominated its products.

On Monday, Openai revealed that changes are being made to promote a healthier relationship between users and Chatgpt.

exist Blog Postsit said: “AI can feel more sensitive and personal than previous technologies, especially for vulnerable individuals suffering from mental or emotional distress.”

It says it won’t give the exact answer to the question, like “Should I break up with my boyfriend?”

Instead, this will “help you think – ask questions, weigh pros and cons”, according to the blog post.

In May, Openai pulled one Seriously criticized updates According to Sam Altman, this makes Chatgpt “too likable”.

exist The latest episode Mr. Altman said that in Openai’s own podcast, he was thinking about how people interact with his products.

“It’s not everything is good, there’s still going to be problems,” he said.

“People will develop these problems, or they may be very problematic, are social relations [with AI]. Society will have to find new guardrails. However, disruption will be huge. ”

As we all know, Altman is a fan of the 2013 movie, and one person has a relationship with his AI peers.

In 2024, actress Scarlett Johansson expressing AI companions in the movie Say she was “shocked” and “irritated” In Openai, the chatbot was launched, and its voice was “exceptionally similar” to her own.