Guest essay published November 9, 2026, argues AI leaders are repeating the virtual reality marketing mistake at far larger scale

Jaron Lanier, the computer scientist who pioneered research in virtual reality, argued in a New York Times guest essay that the technology industry is damaging public acceptance of artificial intelligence through the way its leaders describe it. "Silicon Valley has a special talent for offering amazing products and then screwing them up with ridiculous ideas about how those products should be used," Lanier wrote. He said he works on AI, uses it and likes it, and that "the latest large models are genuinely useful."

Public disapproval is a reaction to messaging, not to the technology

Lanier's central claim is that American disapproval of AI is "unfortunate and unnecessary" and "does not arise from the technology itself." He pointed to the results of the most recent election, which he said showed that many Americans now disapprove of AI "more than almost anything." In his view, the cause is "the insular culture surrounding A.I., a culture that is dictating how this innovation is packaged, how it is sold and how it is understood."

He wrote that many leaders at the biggest AI companies are asking Americans to embrace the belief that the technology "may kill us and our children, or at the very least drain us of all potential for livelihood and meaning," rather than focusing on how the tools could improve lives. "Americans are not rejecting A.I.," he wrote. "They are rejecting the way Silicon Valley is defining it for them."

Anthropic's IPO filing cited as an example

Lanier singled out Anthropic, which he described as the leading AI lab, noting that its filing for an upcoming IPO valued at more than $1 trillion warned that its models could present "catastrophic or existential risks." He offered this as an illustration of how risk framing from the industry's own leaders shapes public perception at a larger scale than earlier technology cycles.

Rhetoric has real-world consequences, including for data center siting

Lanier argued that language filters into behavior. He cited the way some in Silicon Valley "callously treat places where we want to build data centers as sacrifice zones." He wrote that technologists are "sabotaging ourselves," but that it is hard to find a sane focus "from inside the intensity and elite isolation of this bubble."

The virtual reality precedent

Lanier, who said he came up with the term virtual reality, drew a direct parallel to the last major platform cycle. When VR went mainstream around 2013, he said, he realized Americans would not embrace it for a generation because of how key tech leaders talked about it. He cited Meta chief Mark Zuckerberg proclaiming that people would primarily live their lives through VR headsets, and venture capitalist Marc Andreessen suggesting that most people were doomed to live virtually while only elites kept the privilege of living in the physical world. "Predictably, the public did not embrace this vision," Lanier wrote. He said the AI version of the problem is "worse — and at a larger scale."

"Tools not creatures" as a design and safety framework

Lanier proposed a framing he said he coined in the 1980s during arguments with his mentor Marvin Minsky, the MIT professor who helped shape AI culture: "tools not creatures." Because there is no definitive evidence that AI is a conscious entity, he wrote, he is "entirely within my rights — as is anyone else within theirs — to think of the technology as more akin to a fancy calculator than, say, a German shepherd."

Lanier said that choice helps in designing more useful products and, as he intends to show in the remainder of the essay, "provides a concrete and achievable way to make A.I. safe." The excerpt available ends before he lays out that safety argument in detail.