Microsoft CEO Satya Nadella on Google’s challenge using AI technology: “It’s a new race”

Over the past two decades, more people have used Google to explore the web than any other search engine. Now Microsoft is trying to challenge that dominance with breakthroughs in artificial intelligence.

Microsoft unveiled an expanded version of its Bing search engine on Tuesday. In addition to the usual search results, ChatGPT-like technology can answer complex questions, help users make decisions, and turn even complex questions into conversational responses.

For Microsoft CEO Satya Nadella, it’s all a generational opportunity to put his company back at the forefront of innovation.

“It’s a new race in the most important software category or the largest software category in the search. Let’s face it,” Nadella told CBS News’ Tony Dokoupil. “Google dominates it. We’re excited to launch Bing here to keep up.”

Microsoft developed the technology in partnership with OpenAI, the research lab in which it has invested billions of dollars. OpenAI is also behind the viral chatbot ChatGPT. The new AI model is touted as more powerful than its predecessor, but in an early demonstration for CBS News, the feature was intermittently slow, unresponsive, and inaccurate.

Nadella said the only way for a new technology to be “really perfected” is to get “real human feedback” in the marketplace.

With AI in particular, “it has to align with human preferences, both personal and societal in terms of norms. And that’s why we want to introduce them,” he said. “We want to have complete security. We want to have all the things that make sure no damage occurs. But we need them out there in the real world.”

Microsoft revamps search engine with AI technology 01:58

Nadella said the model was certainly trained as a top priority and it will not help anyone do anything illegal.

“We’re going to have many, many mechanisms to ensure that nothing biased, nothing harmful is generated,” Nadella said.

A Microsoft executive declined CBS News’ request to test some of these mechanisms, stating that the functionality in the version used for the demonstration was “probably not the best thing”.

Nadella also addressed concerns about “runaway AI,” which he says would pose “a real problem” if that happened.

“The way to deal with that is to make sure it never runs away,” he said.

“And so I look at it and say…let’s start with…the context in which AI is being used,” Nadella said. “The first categories we should use these powerful models in are where people are clearly and unquestionably in charge. And as long as we start characterizing these models, making these models safer and a lot more explainable over time, then we can think about other uses.”

“But let’s not let it run away,” he said.

Trending News