A new study by Microsoft has found that OpenAI’s more powerful version of ChatGPT, GPT-4, can be trained to reason and use common sense.
Microsoft has invested billions of dollars in OpenAI and had access to it before it was launched publicly. Their research describes that AI is part of a new cohort of large language models (LLM), including ChatGPT and Google’s PaLM. LLMs can be trained in massive amounts of data and fed both images and text to come up with answers.
The Microsoft team has recently published a 155-page analysis entitled “Sparks of Artificial General Intelligence: Early experiments with GPT-4.” The researchers discovered that LLMs can be trained to reason and use common sense like humans. They demonstrated GPT-4 can solve complex tasks in several fields without special prompting, including mathematics, vision, medicine, law and psychology.
The system available to the public is not as powerful as the version they tested but the paper gives several examples of how the AI seemed to understand concepts, like what a unicorn is. GPT-4 drew a unicorn in a sub programming language called TiKZ. In the crude “drawings”, GPT4 got the concept of a unicorn right. GPT-4 also exhibited more common sense than previous models, like ChatGPT, OpenAI said. Both GPT-4 and ChatGPT were asked to stack a book, nine eggs, a laptop, a bottle and a nail.
While ChatGPT recommended placing the eggs on top of the nail, the more sophisticated model arranged the items so the eggs would not break.
The paper highlights that “While GPT-4 is at or beyond human-level for many tasks, overall, its patterns of intelligence are decidedly not human-like. However, GPT-4 is almost certainly only a first step towards a series of increasingly generally intelligent systems, and in fact, GPT-4 itself has improved throughout our time testing it.”
However, the report acknowledged that AI still has limitations and biases and users were warned to be careful. GPT is “still not fully reliable” because it still “hallucinates” facts and makes reasoning and basic arithmetic errors.
[Link to paper: Sparks of Artificial General Intelligence:Early experiments with GPT-4]
Additional Information
Samuel Altman, the chief executive of company OpenAI that owns artificial intelligence chatbot ChatGPT, testified before the United States Congress on the imminent challenges and the future of AI technology. The oversight hearing was the first in a series of hearings intended to write the rules of AI.
[Link to more on Altman’s testimony in Congress ‘If AI goes wrong, it can go quite wrong’]