Microsoft, OpenAI admit to core problems internally and with users

According to reporting by The Information, Satya Nadella noted that Microsoft programs for connecting Copilot with Gmail and Outlook “for the most part don’t really work” and are “not smart.”

Furthermore, Nadella has told colleagues that Office 365 isn’t delivering on its promise to meaningfully automate work. Some large customers aren’t using the AI very much.

Microsoft needs to convince corporate customers that 365 Copilot is more useful than free or cheaper paid alternatives such as ChatGPT or Google’s Gemini. Nadella has privately voiced concerns that Copilot isn’t as good as Gemini at connecting to other applications to carry out complex tasks.

To that end, Microsoft is now paying third parties tens or hundreds of thousands of dollars to train users on Copilot.

Microsoft isn’t the only AI provider whose customers don’t know how to use the AI. OpenAI has similarly found that ChatGPT users also are pretty bad at using the chatbot to its fullest potential.

Even as ChatGPT attracted more users this year, improvements to the underlying AI model’s intelligence—and the in-depth research or calculations it could suddenly handle—didn’t seem to matter to most people using the chatbot, who are probably asking about pretty simple things, like movie ratings, where you wouldn’t need a model to think for half an hour.

In other words, there seems to be no aparent, measurable change in use cases after major upgrades.

The change in the way ChatGPT users have reacted to new models powering the chatbot shows how the goals of OpenAI’s core AI research division, which develops its technology, don’t always serve the needs of ChatGPT.

Divisions in the ranks

OpenAI’s head of applications, Fidji Simo, wrote on her blog that at its core, OpenAI remains a research-focused company and that “products aren’t the goal themselves,” despite the fact that ChatGPT drives most of the company’s revenue.

There are other signs of schisms between research and product at OpenAI. When Google released its image-generating AI, Nano Banana, in August, there was a move within OpenAI to improve its own image technology.

This prompted a disagreement between Altman, who felt image-generation capabilities were important to ChatGPT’s growth, and Mark Chen, the company’s research chief, who wanted to prioritize other initiatives.