Microsoft’s Satya Nadella Wants Us to Stop Treating AI Like Humans

(Bloomberg) -- A week after OpenAI unveiled a personal assistant that can laugh, sing and speak with a mix of different voices, the company's closest partner offered a subtly different view for how people should engage with artificial intelligence tools. 

“I don't like anthropomorphizing AI,” Microsoft Corp. Chief Executive Officer Satya Nadella told Bloomberg Television on Monday, referring to the practice of using verbs and nouns to describe AI that are typically reserved for people. “I sort of believe it's a tool.”

Nadella’s remarks hint at an ongoing debate in the tech industry over how much to humanize AI services at a time when the technology is advancing and responding in ways that appear more human-like. Last week, a Google executive told Bloomberg that while it's possible to build AI tools that “show emotion,” the company prefers to focus on “being super helpful and super useful.”

OpenAI has taken a different approach. The company demonstrated a new voice assistant last week that it said can understand emotions and express feelings of its own. At multiple points in the presentation, the AI voice appeared to hit on the employee using the tool onstage. Many on social media likened the feature to the dystopian movie “Her,” a comparison fueled by one particular voice option that users said resembled the film’s star, Scarlett Johansson. 

Johansson said in a statement given to NPR that OpenAI CEO Sam Altman reached out and asked her to consider voicing an audio chat feature. According to Johansson, Altman tried to pitch her on the idea that she could “help consumers to feel comfortable with the seismic shift concerning humans and AI.” She declined and said she has since been forced to hire lawyers over OpenAI's decision to move forward with a similar sounding voice. (OpenAI has since taken down the voice and replaced it with another.)

Even before ChatGPT brought AI into the mainstream consciousness, tech companies often conferred human personalities onto AI programs, usually with female-coded names and characteristics, in an apparent effort to help people connect and feel comfortable with the technology. Nadella’s Microsoft hasn’t been immune to that behavior either. The company has released various conversational and AI programs over the years, including Tay and Cortana, named after Halo’s female-appearing AI assistant. And who can forget Bing AI’s rogue persona, Sydney?

There’s a natural tendency to want to describe artificial intelligence in human terms, as people look to explain the math, numbers and code behind the software in ways that users can relate to, saying things like AI “learns.” That temptation will only get stronger as tech companies release more capable products that can hold real-time conversations. 

But in the interview, Nadella said users need to be mindful that the abilities AI software displays are not human intelligence. “It has got intelligence, if you want to give it that moniker, but it’s not the same intelligence that I have,” he said.

In fact, Nadella went so far as to lament the selection of the term “artificial intelligence,” first coined in the 1950s. “I think one of the most unfortunate names is ‘artificial intelligence’ — I wish we had called it ‘different intelligence,’” he said. “Because I have my intelligence. I don’t need any artificial intelligence.”

Nadella just wants the AI software to help when he wants it. “That’s, I think, the ideal relationship.”