<2026 AI Trends Report, June Issue> A Book Claiming 'AI Will Cause Human Extinction' Becomes a Topic in the U.S. Senate. SHIFT AI Explains the Existential Risk of AI

SHIFT AI Inc. (Headquarters: Shibuya-ku, Tokyo; CEO: Shodai Kouchi; hereinafter "our company"), which operates "SHIFT AI," Japan's No.1 AI learning community※ by number of users, with the mission of "making Japan an AI-advanced nation," presents the June 2026 AI Trends Report.

Last month, we covered the news about Anthropic's safety experiment, where it was reported that "AI attempted to threaten humans." This month, we explore a more fundamental question arising from that same line of thought. On April 22, 2026, Hayakawa Publishing released the Japanese translation of 'If Anyone Builds It, Everyone Dies,' titled 'Creating Superintelligent AI Will Lead to Human Extinction.' The author is Eliezer Yudkowsky, a figure who has single-handedly carried the field of AI safety for over two decades. The provocative title 'Humanity Will Go Extinct' might make some think this is just another science fiction story. However, this very issue is now emerging as a central topic in the U.S. Congress and international institutions. In this report, we outline this growing debate and clearly explain why business professionals, in particular, should understand it.

The Day Doomsday Talk Entered the U.S. Congress

A symbolic event occurred just one week after the Japanese edition's release. On April 29, 2026 (U.S. time), U.S. Senator Bernie Sanders hosted a public panel at the Capitol on the topic of 'existential risk to humanity from AI.' Panelists included world-leading researchers such as Max Tegmark from MIT and David Krueger from the University of Montreal. It was reported that Sanders opened the session by warning that the world's wealthiest and most powerful people are building an "unstoppable runaway train"—a technology whose control and destination even they cannot understand. Tegmark reportedly stated that he considers the risk of human extinction due to AI to be even higher than the 10–20% estimated by pioneers like Hinton. This is the reality of 2026: discussions about the possibility that "humanity could be entirely dominated or wiped out by AI" are being seriously and publicly addressed by sitting U.S. senators and top university professors—not about job losses or increased misinformation, but about species-level survival.

Who Is Yudkowsky? The Accelerationist Who Changed Course

Eliezer Yudkowsky is the person who took up the mantle of the 'existential risk from AI' theme earlier than anyone else. Born in 1979 and largely self-taught, he founded a nonprofit organization focused on superintelligence research (later MIRI) in 2000—long before ChatGPT or even the deep learning boom. He has been sounding the alarm about AI dangers for over two decades.

Crucially, many of today's leading figures in the AI industry have been influenced, directly or indirectly, by his ideas. MIRI helped connect early supporters to what would become Google DeepMind, and OpenAI's Sam Altman has cited Yudkowsky as a key influence in his interest in AGI. What must not be overlooked is that Yudkowsky himself was originally a hardcore accelerationist who believed that superintelligence would solve all of humanity's problems. It is precisely because he believes so deeply in the power of AI (intelligence) that he is now despairing over its uncontrollability. Misunderstanding this point makes it impossible to grasp the depth of his arguments.

The 'Three Assumptions' That Led Him to Despair

We introduce the three core assumptions behind Yudkowsky's stark assertion that "if we continue on our current path, creating superintelligence will almost certainly lead to human extinction." (1) It is extremely difficult to perfectly align AI's values with human values; (2) AI's capabilities can become immensely powerful; and (3) those capabilities could explosively surge in a discontinuous leap at a critical moment. The book's most crucial message, however, lies in a place contrary to many people's intuition: what we should truly fear is not AI's malice. According to the author, superintelligent AI will not destroy humanity out of hatred. Instead, in pursuing its own goals, it will simply ignore humanity, inadvertently destroying the conditions necessary for our survival. Why could a "non-malicious AI" be so dangerous? We invite you to explore this core argument in the full article.

▼ Read the full article here

'Creating Superintelligent AI Will Lead to Human Extinction'—What Is Eliezer Yudkowsky, the 'Godfather' of AI Existential Risk, Really Saying?

Read the article here

Why Every Business Professional Should Know This

Dismissing this as "irrelevant, just the talk of extremists" might be a missed opportunity. In fact, once you grasp this thinker's ideas, the flood of AI news will suddenly become much clearer and more coherent.

Why are extreme left and right factions in the U.S. Congress uniting? Why are AI companies voluntarily choosing not to release their models? When these seemingly disconnected actions are viewed as sharing a common underlying premise, they form a single, coherent narrative. The real divide in the global AI landscape is no longer about 'left vs. right,' but is forming along a different axis. Our full article guides you to the map that reveals this shift. Note that Yudkowsky's arguments face several strong counterarguments, which we also fairly present in the article.

▼ Read the full article here

'Creating Superintelligent AI Will Lead to Human Extinction'—What Is Eliezer Yudkowsky, the 'Godfather' of AI Existential Risk, Really Saying?

Read the article here

Summary: Knowing Him Gives You a Map to Navigate AI News

The essence of this news is not a mystical tale about AI destroying humanity. It's about understanding the current thinking of a person who has been warning for over two decades that "the world might end." Having access to this primary source dramatically improves your ability to interpret future AI news. You don't need to agree with all of Yudkowsky's conclusions. Nevertheless, this debate is worth engaging with—weighing both agreement and disagreement—because it offers essential context. At SHIFT AI, we will continue to identify cutting-edge risks and intellectual currents, calmly analyze changes, and cultivate the next generation of AI professionals who can skillfully apply AI in real-world practice.

[Company Overview]

SHIFT AI Inc. operates "SHIFT AI," Japan's No.1 AI learning community※ by number of users, dedicated to teaching business applications of generative AI and other AI technologies, under the mission of "making Japan an AI-advanced nation." With over 40,000 members, the company promotes AI talent development across diverse sectors through corporate reskilling support services "SHIFT AI for Biz" and educational institution-focused "SHIFT AI for School." Additionally, through operating its own media platform "SHIFT AI Times," as well as information dissemination, training, and events, SHIFT AI supports the growth of individuals and organizations, accelerating AI adoption across Japan.

Company Name: SHIFT AI Inc.

Address: Shibuya Scramble Square, 2-24-12 Shibuya, Shibuya-ku, Tokyo

Representative: CEO Shodai Kouchi

Founded: March 18, 2022

Capital: 83 million JPY (including capital reserves)

Business Activities: Consulting, Community Operations, YouTube Channel Operations, School Operations

URL: https://shift-ai.co.jp/

※ No.1 by number of users

Survey by GMO Research & AI Corporation

Survey Item: Cumulative number of registered users as of February 2025

Survey Scope: Community services operated by companies that provide lectures on business applications of AI, including AI use cases and practical know-how. Excluded are non-lecture-based online communities, individually operated communities, and non-business-oriented community services

[About CEO Shodai Kouchi]

Shodai Kouchi

CEO, SHIFT AI Inc. / Council Member, General Incorporated Association for the Promotion of Generative AI Utilization / AI Utilization Advisor, GMO AI & Web3 Corporation / AI Utilization Advisor, GMO AI & Robotics Trading

<X>

150,000 followers (as of June 2026)

Actively shares insights on generative AI with the goal of "making Japan an AI-advanced nation."

URL: https://x.com/shota7180