A 3D render of the Google Assistant app icon on a blue background
An arrow pointing leftHome

Google CEO says their crappy chatbot proves AI's not sentient

  • Justin Caffier

So much for the LaMDA controversy?

Google CEO Sundar Pichai loves making news. Sure, as the leader of one of the biggest companies in the world, the bulk of the job does revolve around figurehead duties. But the amount of headlines the man has generated over the past few weeks alone by merely making statements to the press belies a level of C-suite clout chasing we haven’t encountered since the last presidential administration.

Just a few of the recent story-generating bombshells by the McKinsey-engineer-turned-newsmaking-tech-deity: a warning about the economy, hints about forthcoming layoffs, why TikTok is so scary and robust a competitor that we should all stop talking about antitrust legislation. With all these stories flitting about the information superhighway, it’s easy to see how the Pichai quote most relevant to our interests almost fell through the cracks.

Speaking at Vox Media’s Code Conference in Beverly Hills earlier this month, on the very same dais where he generated so many of these other headlines, Pichai assured the audience that, not only was Google’s AI LaMDA chatbot “not sentient by any stretch of the imagination,” but the $1.4 trillion company with offices around the globe—including major hubs in Seattle and Portland—may never produce AI technology that passes the Turing test.

Pichai’s downplaying of the company’s AI prowess was meant to quell concerns about the recent news controversy stirred by former Google employee Blake Lemoine, who was put on administrative leave after suggesting LaMDA had gained sentience akin to that of a 7-to-9-year-old child and thus deserved the same rights and legal representation as any human.

Claiming he was surprised by how much traction the Lemoine incident got in the press, Pichai went on to assure the audience that Google’s AI still has “a long way to go” and that it’s easy for him to get bogged down in “philosophical or metaphysical talks often about what is sentience and what is consciousness.”

Hoping to underscore his refutation of Lemoine’s claims with a bit of humor, Pichai pointed to a notoriously wonky Google product as evidence of the state of their AI development.

“The good news,” joked Pichai, “is that anyone who talks to Google Assistant—while I think it is the best assistant out there for conversational AI—you still see how broken it is in certain cases.”

Despite the brushed aside sentience concerns, Pichai made clear Alphabet and Google’s commitment to and reliance on AI now and going forward. He emphasized that the foundational human values driving the company’s AI ambitions are more important and what will imbue them with the potential to “be empowering to humanity.”

Whether or not they’re in earnest, Pichai’s words are just the sort of thing AI acolytes are hoping to hear these days. As the technology has proliferated thanks to Alphabet and other giants, its most glaring flaws—like its penchant for pro-white racial bias—have revealed themselves and taught us all that AI is only as innocent and egalitarian as the people and society it’s a product of.

As for Pichai, there’s no denying the guy could probably handle press conferences, congressional testimonies, and “casual” seated chats with other industry leaders all day every day. He seems to be almost tailor-made for addressing the latent concerns of a global population eager to use his products but desperate for reassurance that their personal values will remain uncompromised when they do and generating press for the company in the process.

Y’know, if one was a tad more conspiratorial-minded, they might start to assume his denial of Google’s AI sentience capabilities coupled with his uncanny ability to pitch layup stories to the media was hinting at something that would put Sundar Pichai at the center of the biggest news story of the century. But, then again, the company’s virtual assistant is pretty bad.