OpenAI, the Microsoft-backed AI startup, has started rolling out GPT-4, its latest artificial intelligence model, setting the stage for a new wave of human-like technology.
The release could intensify the rivalry between Microsoft Corp, which backs OpenAI, and Alphabet Inc’s Google.
OpenAI announced in a blog post that GPT-4 is “multimodal,” meaning it can generate content from both text and image inputs.
The text-input feature will be accessible to ChatGPT Plus subscribers and developers on a waitlist, while the image-input function is still in a research preview phase.
This launch suggests that office workers might rely even more on AI for a broader range of tasks, highlighting the ongoing competition among tech giants to capitalize on these advancements.
Alphabet’s Google also recently introduced a new AI tool for its collaboration software that can draft nearly any type of document, just ahead of Microsoft’s plans to showcase AI enhancements for its Word processor, likely utilizing GPT-4.
A Microsoft executive confirmed that GPT-4 is already helping to power Bing’s search capabilities.
GPT-4 offers significant improvements over its predecessor, GPT-3.5. For instance, in a simulation of the bar exam for U.S. law school graduates, GPT-4 scored in the top 10% of test takers, whereas GPT-3.5 ranked around the bottom 10%.
While GPT-3.5 and GPT-4 may appear similar in casual conversation, OpenAI states that “the difference comes out when the complexity of the task reaches a sufficient threshold,” noting that GPT-4 is “more reliable, creative, and able to handle much more nuanced instructions.”
During an online demonstration, Greg Brockman, OpenAI’s president, showed that GPT-4 could turn a hand-drawn sketch into a functional website and even assist users in calculating their taxes.
Sam Altman, OpenAI’s CEO, described GPT-4 on Twitter as the company’s “most capable and aligned” model with human values and intent, but he acknowledged that “it is still flawed.”
According to OpenAI, GPT-4 is 82% less likely to respond to requests for disallowed content compared to GPT-3.5 and scores 40% higher on specific tests for factual accuracy, addressing the common AI challenge of generating inaccurate information or “hallucinations.”
Rishi Jaluria, an analyst at RBC Capital Markets, notes that Microsoft is well-positioned to benefit from GPT-4’s adoption.
Microsoft is integrating GPT-4 into its products, and its Azure cloud service is the backbone for OpenAI’s technology, which comes at a critical time when companies are carefully managing IT spending amid economic uncertainty.
“Whenever a company uses this piece of technology, those workloads go through Microsoft Azure,” Jaluria emphasized, highlighting the strategic advantage Microsoft gains from GPT-4’s rollout.
