Microsoft AI researchers published a paper claiming that the OpenAI language model powering Microsoft’s Bing AI exhibits “sparks” of human-level intelligence, or artificial general intelligence (AGI).
GPT-4’s abilities were described by the researchers as “only a first step towards a series of increasingly generally intelligent systems,” rather than fully formed, human-level AI. They add that it outperforms previous OpenAI models in novel and generalized ways. GPT-4 can solve novel and difficult tasks in mathematics, coding, vision, medicine, law, psychology, and other fields without any special instruction.
“Moreover, in all of these tasks, GPT-4’s performance is strikingly close to human-level performance, and often vastly surpasses prior models such as ChatGPT,” reads the paper.
However, the researchers admitted in the paper that while GPT-4 is “at or beyond human-level for many tasks,” its overall “patterns of intelligence are decidedly not human-like.” The researchers also addressed that AGI still doesn’t have a firm, agreed-upon definition, and neither does the more general concept of “intelligence.” The paper claimed that “GPT-4’s intelligence signals a true paradigm shift in the field of computer science and beyond.”
The paper also stated that GPT-4’s performance is based on an early version of the language model and is not necessarily representative of the product-applicable version.
The sources for this piece include an article in Futurism.