What do we really know about GPT-4 and what we can learn about it, and how we can use it to improve our understanding of artificial intelligence?
Bubeck and other AI researchers at Microsoft were inspired to wade into the debate by their experiences with GPT-4. A few weeks after Bing’s new chat feature launched, the company released a paper that claimed GPT-4 had “the potential to be an Artificial General Intelligence” according to the paper.
Bubeck went to his computer, and asked GPT-4 to use a programming language called TikZ to create a scientific diagram. Bubeck was using a version of GPT-4 that only worked with text, not images. But the code the model presented him with, when fed into a TikZ rendering software, produced a crude yet distinctly unicorny image cobbled together from ovals, rectangles, and a triangle. Such a feat had to be grasped by an abstract grasp of the elements of the creature. Something’s happening that he says is new. “Maybe for the first time we have something that we could call intelligence.”
Having a clear grasp of what new abilities are and what are not is important for understanding the potential or risks. While there is broad agreement that the systems give computers significant new skills, researchers are just beginning to study the behaviors and determine what is behind the prompt.
While Openai promotes GPT-4 as a good bet for bar and med school exams, scientists who study aspects of human intelligence say its remarkable capabilities differ from our own. The models’ tendency to make things up is well known, but the divergence goes deeper. And with millions of people using the technology every day and companies betting their future on it, this is a mystery of huge importance.
My Artificial Intelligence: A Discovery-Based Service to Help People Make AdSense of Machine Learning Technologies Is Open Questions and Answers to Human Concerns
The company is giving all 750 million monthly users of its app a free version of its My Artificial Intelligence bot, just two months after it was released to the 3 million paid subscribers.
“Just based on the way that they work, I think they’re much more suited to creative tasks,” Spiegel says of generative AI bots. “And some of the things that make them so creative are also the things that make them not so great at recalling specific information.”
Spiegel is tight-lipped about the impact of my artificial intelligence on the advertising business. He acknowledges that leveraging My AI’s interactions for ad targeting could be an opportunity but refrains from elaborating further, hinting at possible developments in the near future.
Despite the problematic interactions that have already surfaced, Spiegel says the overwhelming majority of interactions with My AI have been positive. “The thing that gave us a lot of confidence is that as we monitored the way that people were using the service, we found that 99.5 percent of My AI replies conformed to our community guidelines,” he says.
There is a debate in the artificial intelligence industry on the issue of whether to make a product more Human-like. It was one of the requests made by early users to change My AI’s name and appearance. It means that the human desire to make things feel like they are their own is what speaks to me.
As for the broader concerns regarding the potential harm of generative AI, Spiegel offers an optimistic perspective: “When I compare this to almost any other technology that has been invented in the last 20 years, it’d be hard to name one where people have been more thoughtful about the way it’s being implemented and rolled out.”