OpenAI unveils ChatGPT successor with ‘human-level’ performance
GPT-4 able to pass bar exam with a score in the top 10 percent of applicants, says the California-based company.
ChatGPT’s long-awaited successor has gone live, showing off “human achievements” in university standard exams.
OpenAI said GPT-4, the next generation of its artificial intelligence-powered chatbot, marked a “milestone” in the development of deep learning, which mimics how humans acquire knowledge.
“We spent 6 months iteratively tuning GPT-4 using lessons learned from our hostile testing program and ChatGPT, resulting in our best-ever results (though far from perfect) on factuality, controllability, and refusal to go outside the guardrails. the San Francisco-based company said in a blog post Tuesday.
OpenAI, which is backed by Microsoft, said the new version of its AI-powered chatbot is a “multimodal” model that can generate content from both images and text prompts.
In an online demonstration, OpenAI President Greg Brockman showed how GPT-4 created a real website from a hand-drawn model.
OpenAI said the update is able to pass the bar exam for prospective lawyers with a score in the top 10 percent of applicants, compared to the bottom 10 percent of test-takers previously.
The chatbot can also beat 90 percent of people who take the evidence-based reading and writing portion of the Scholastic Assessment Test and the verbal portion of the Graduate Record Examination used for admission to postgraduate education, according to OpenAI.
GPT-4 is also much less likely to produce inaccurate, offensive, abusive responses than ChatGPT, the company said.
“We spent six months making GPT-4 more secure and better tuned. GPT-4 is 82 percent less likely to respond to requests for disallowed content and 40 percent more likely to produce factual responses,” according to OpenAI.
However, OpenAI said GPT-4 is still “not completely reliable” and may still produce unexpected responses known as “hallucinations” and errors of reasoning.
OpenAI’s launch of ChatGPT in November has taken the tech world by storm, raising existential questions about the future of industries ranging from education to journalism and healthcare.
Tech giants including Google, Microsoft, Huawei, Alibaba and Baidu are racing to roll out their own versions of the technology amid heated competition to dominate the burgeoning AI sector.