Shan Oppa, Golden Finance

In the spring of 2026, the gunfire in San Francisco’s Pacific Heights community shattered a generally accepted assumption: opposition to artificial intelligence may no longer be limited to denunciations on keyboards and protests behind screens.

When the flames of Molotov cocktails cut through the night sky, and when bullets penetrated the door and window glass, pointing directly at the residence of OpenAI CEO Sam Altman, we had to face a cruel reality: in the shadow of the rapid iteration of the technological revolution, human beings' fear of AI has been transformed into a weapon of harm.

One, 48 hours, two shocking attacks

At 2:30 in the morning on April 11, 2026, in the quiet Pacific Heights neighborhood of San Francisco, a Molotov cocktail suddenly pierced the dark night and hit the outer wall of Sam Altman's residence. The flames briefly burst out, causing only slight damage to the exterior wall of the house and no casualties, but it was enough to shock the entire technology community. It is reported that after committing the crime, the three masked suspects quickly boarded a black Honda Civic and fled the scene in a hurry.

Just 48 hours later, panic struck again. At 1:15 a.m. on April 13, at the same place, a car slowly stopped, followed by a gunshot that broke the silence, and the window glass of the residence shattered. Fortunately, there were still no casualties in this attack. The police quickly intervened and arrested two suspects on the spot - 25-year-old Amanda Tom and 23-year-old Muhamad Tarik Hussein. The two were detained according to law for "negligent discharge of firearms". At present, the Federal Bureau of Investigation (FBI) has officially intervened in the investigation, and the case will most likely be upgraded to terrorism-related charges.

Subsequently, an anti-AI group called "Pause AI" publicly claimed responsibility for both attacks. This is not a Hollywood suspense script, but a real violent incident that happened to core figures in the technology circle in the spring of 2026.

2. From peaceful petitions to violent attacks: the out-of-control escalation of the anti-AI movement

If we go back two years, the anti-AI movement still remains in the category of rational protests and peaceful petitions, full of restraint and expression of demands.

In September 2024, the "AI Safety Network" was officially established. Its core activities focus on policy lobbying, public petitions and exposure of industry chaos, trying to promote the implementation of AI supervision through legal channels. In the same year, hundreds of anti-AI people gathered outside the OpenAI headquarters, holding high the slogan "Suspend AI R&D", calmly but firmly conveying their demands to the government and enterprises, calling for the establishment of a complete AI regulatory system.

The turning point occurred in 2025. The direction of the anti-AI movement completely changed, and moderate demands were gradually replaced by radical actions.

In August of this year, more than 500 protesters surrounded the OpenAI headquarters. The order at the scene was out of control. Three people were arrested by the police for disturbing public order. In November, a Google data center in Oregon was deliberately sabotaged by anti-AI activists, causing a power outage for up to 2 hours and causing direct economic losses of up to US$2 million. This is the first time that an anti-AI action has resulted in substantial property damage, and it also marks the movement’s official turn to radicalization.

As we enter 2026, the anti-AI movement has further accelerated its escalation, and its violent tendencies have become increasingly obvious.

In January, more than 120,000 people signed a petition asking the White House to urgently suspend the GPT-5 training program, fearing that it might bring uncontrollable risks; in February, a large-scale sit-in demonstration broke out outside the office of Anthropic in New York, and 12 protesters were arrested for refusing to cooperate with the evacuation; in March, the NVIDIA data center in California was damaged again, the loss expanded to 5 million US dollars, and two suspects were arrested; in April, anti-AI people directly targeted AI The core figure in the industry, Sam Altman, used Molotov cocktails and bullets to push the protests to the peak of violence.

From signature petitions to vandalism, from property attacks to personal threats, the escalation trajectory of the anti-AI movement is strikingly similar to the historical Luddite movement—in 1811, British textile workers worried that machines would take away their jobs, so they smashed textile equipment and used extreme methods to resist technological change. History will not simply repeat itself, but it will always echo in the course of time with similar rhythms, which also lays a historical footnote for the current AI controversy.

3. Six major anxieties behind anti-AI sentiment

To understand the out-of-control anti-AI movement, we must first understand the deep-seated fears behind it. These fears are not groundless, but the real impact that technological change has on human society, living conditions, and self-worth.

1. Jobs are replaced

This is the most direct and common fear. Data from 2025 shows that about 3.8 million people in the United States have lost their jobs due to AI applications, including customer service, translation, junior programming, data analysis and other white-collar positions that were once considered "stable iron rice bowls." These positions are being replaced in batches by efficient AI systems. McKinsey predicts that by 2030, about 45% of white-collar jobs worldwide will be directly affected by AI and will either be replaced or forced to transform.

A former claims adjuster who has worked for an insurance company for 15 years admitted in an interview in November 2025: "I have been doing claims review for 15 years. Last year, out of 80 employees in the department, only 5 people were left to deal with 'abnormal cases' that cannot be solved by AI. The remaining 75 people were all replaced by AI systems. We don't even have much buffer time." For enterprises, the adoption of AI can reduce the average cost. 60%, shareholder returns have increased significantly, but for ordinary employees, this means a decrease in income and a sharp increase in the risk of unemployment. The improvement in efficiency has not been simultaneously translated into employment growth. This is the cruelest reality behind technological change.

2. Losing control

Human beings are born with a fear of things that are “incomprehensible and uncontrollable”. No matter how complex traditional technology is, humans can grasp its core principles and can intervene, adjust or even stop operation at any time. However, the core of AI technology is "black box decision-making", and its operating logic and judgment basis are difficult to be fully interpreted by humans. It may even develop autonomous decision-making capabilities during the learning process, which exceeds human expectations and control.

As Stuart Russell, a professor at the University of California, Berkeley, said: "The most terrifying misunderstanding that humans have about AI is not that it will rebel against humans, but that it will perfectly execute humans' wrong instructions - when we do not consider all risks, AI's 'absolute obedience' may have devastating consequences."

3. Rising social inequality

The technological revolution is supposed to bring universal benefits, but the distribution of AI dividends has shown serious polarization. Relevant estimates for 2025 show that of the total value created by AI, shareholders of top AI companies will take away about 45%, and core AI researchers and engineers will share about 25%, while ordinary workers will only receive about 15% of the benefits. A large number of unemployed people replaced by AI will not even receive any dividends.

A protester from the "Pause AI" organization said angrily during a demonstration in February 2026: "The CEOs and shareholders of AI companies are making a lot of money, making billions a year, but we are losing jobs and livelihoods because of AI. This is not technological progress at all, this is naked plunder."

4. The disappearance of human uniqueness

Art, writing, thinking, and creation, these abilities that were once considered "only for humans" are now being broken through by AI one by one. The paintings generated by Midjourney have won international awards, GPT-5 has successfully passed the top literary test, and Devin can independently complete complex programming projects. When AI can be better at creating, better at thinking, and even better at understanding human emotions and needs than humans, many people begin to be confused: What is the unique value of human existence?

5. Privacy leak

The popularization of AI technology has also brought about an unprecedented privacy crisis. Currently, 85% of major cities around the world have deployed AI facial recognition systems, and 90% of Internet user data has been collected and analyzed by AI companies. The accuracy of AI's prediction of personal behavior has reached as high as 78%. Our travel trajectories, consumption habits, chat records, and interest preferences are all being quietly captured and interpreted by AI. Human beings seem to live in a transparent world without privacy, and their sense of security is constantly being eroded.

6. Doomsday Panic

The media is keen to exaggerate the doomsday scenario of "AI out of control and human extinction", exacerbating public panic. However, according to the assessment of industry experts, the probability of occurrence of these extreme scenarios is extremely low: mass unemployment (85%, has occurred), deepfake abuse (75%, has occurred), AI weaponization (45%, expected to occur in 5-10 years), AI completely out of control (5-15%, expected to occur in 10-30 years), and human extinction (less than 1%). Unfortunately, the media often exaggerates the low-probability risks of doomsday, while ignoring the real harm that has already occurred and has wide-ranging impacts.

4. AI controversy in the eyes of celebrities

Faced with the opportunities and risks brought by AI, celebrities from all walks of life around the world have divided their views, forming three distinct camps, each of which stands firm and engages in fierce debates.

Opposition: Resolutely call for a halt and be alert to risks

This camp has the loudest voice, and its core demand is to suspend AI research and development to prevent potential risks. Elon Musk publicly stated in December 2025: "AI is the biggest threat facing human civilization. If it is not controlled in time, the consequences will be disastrous." Apple co-founder Steve Wozniak also said in January 2026: "The development speed of AI has exceeded human control, and it must be paused immediately and reassessed its risks." Yuval Harari, author of "A Brief History of Humanity", warned in September 2025: "AI It may completely end the human era and reshape the ecology and order of the entire earth.”

The most thought-provoking is Jeffrey Hinton, known as the "Godfather of AI". He resolutely resigned from his position at Google in June 2025 and publicly confessed: "I regret participating in AI research and development. I realize that the technology I created with my own hands may cause irreparable harm to mankind."

Supporters: Embrace change and stick to progress

Supporters firmly believe that AI is an inevitable trend of human progress, and fear should not be a reason to hinder technological development. Sam Altman responded after the attacks in April 2026: "People's fear of AI is reasonable, and I completely understand it, but we cannot stop progress because of fear. Historically, every technological revolution is accompanied by doubts and fears, but in the end it promotes the prosperity and development of human society."

Meta CEO Mark Zuckerberg said in January 2026: "AI is not a threat, but a tool to liberate human creativity. It can help us get rid of tedious repetitive work and focus on more valuable creations." NVIDIA CEO Jensen Huang said bluntly in December 2025: "AI is the core of a new round of industrial revolution. It will reshape the global industrial landscape and bring unprecedented development opportunities."

AI expert Andrew Ng's view is more rational: "The AI threat theory has been over-exaggerated by the media. In fact, AI will not replace humans. What will really be replaced are those who are unwilling to learn and refuse to adapt to AI." a16z founder Marc Anderson is firm: "AI will save the world, and it can solve major problems facing mankind such as climate change and disease prevention. Opposing AI is essentially opposing human progress."

Neutral: Balancing innovation and security

The centrists try to find a balance between "promoting innovation" and "preventing risks", emphasizing the importance of supervision. Bill Gates said: "AI not only brings huge development opportunities, but also lurks risks that cannot be ignored. The key is not 'whether to develop AI', but how to guide the healthy development of AI - we need a sound regulatory system that can neither stifle innovation nor allow risks to spread."

Google CEO Sundar Pichai also holds a similar view: "The core of AI development is balance. It is necessary to maintain the vitality of innovation, but also to maintain the bottom line of safety, so that AI can truly serve human beings." Apple CEO Tim Cook emphasized: "AI should adhere to 'people-oriented', and all technology research and development should focus on human needs and well-being, rather than purely pursuing technological breakthroughs."

5. Mirror of History

The anti-AI movement is not the first rebellion against new technology in human history. Looking back on the past, every major technological revolution has been accompanied by fierce opposition, and the laws of history may provide enlightenment for the current AI controversy.

From 1811 to 1816, the famous Luddite movement broke out in Britain. Textile workers, worried that machines would replace manual labor and take away their jobs, smashed textile equipment and used violence to resist technological change. In the end, the government suppressed the movement, but it also realized the employment problems caused by technological changes, and subsequently introduced a series of labor laws to protect workers' rights and interests. The development of technology is unstoppable, but humans can reduce the harm caused by it through system design.

The rise of railroad technology in the 1830s led to the anti-railway movement. People are worried that trains will destroy farmland, scare livestock, and even believe that "high-speed movement will harm human health" and oppose railway construction. But in the end, with its advantages of efficiency and convenience, the railway gradually became popular and became the core driving force of the industrial revolution, and the anti-rail movement also dissipated.

Since the 1950s, the anti-nuclear movement has continued. Opponents are worried about the devastating consequences of nuclear energy leakage and nuclear war. This movement has achieved partial success, with global nuclear energy development being strictly regulated and the nuclear safety system being continuously improved. Similarly, the anti-GMO movement has never stopped since the 1990s, and Europe has even completely banned genetically modified crops, reflecting mankind's cautious attitude towards the risks of new technologies.

Today, the anti-AI movement from 2024 to the present is still continuing to ferment, and its outcome is still unknown. But history has already given a rule: in the early stage of the emergence of new technologies, they will inevitably encounter fierce opposition; in the mid-term, the government and society will intervene in supervision and standardize technological development; in the later stage, as the technology becomes popular and people gradually adapt, the opposition will gradually weaken.

However, it is worth noting that in every technological revolution, there will be "beneficiaries" and "victims". Some people gain opportunities due to technological progress, and some people lose their livelihood due to technological changes. The core issue has never been "whether to develop technology or not", but "how to fairly distribute technological dividends" and how to protect those harmed by technology. This is also the core demand of the anti-AI movement, and it is a topic that we must face directly at the moment.

六、结语

The bullet holes in Sam Altman’s window are not a threat to one person, but a warning to the entire era.

Those who threw Molotov cocktails were not all lunatics. They include truck drivers who have lost their jobs, graduates who cannot see their future, and parents who are afraid that their children will be replaced by AI. The fears are real and the demands are worth hearing, but violence is not the answer.

The real answer is not in the bullets, but in the system, in the distribution, and in the compensation for the "left behind". History has repeatedly proven that technology itself is neither good nor evil. What is the value of humans when AI can do most of what humans can do, or even better?

References

1. U.S. AI-related unemployment data: Comprehensive Challenger, Gray & Christmas (2025) AI-related layoff statistics, MIT + Oak Ridge Laboratory (2025.12) study of the impact of AI on the U.S. labor force, the "3.8 million people unemployed due to AI" in the article is an estimate of the combined direct + indirect impact.

2. McKinsey forecast data: From the 2024-2025 series of reports from the McKinsey Global Institute (MGI), the article states that "by 2030, approximately 45% of white-collar jobs worldwide will be directly affected by AI", which is the general middle value of the range of "40%-70% of jobs/tasks are affected by AI" in the report.

3. AI value distribution data: Based on consensus research on the AI dividend distribution pattern by academics/think tanks, the "shareholders 45%, core R&D 25%, and ordinary workers 15%" in the article are a quantitative approximation of the industry distribution trend.

4. Enterprise AI cost data: From industry surveys by consulting companies such as Gartner and McKinsey, "the average cost of enterprise adoption of AI has dropped by 60%" is the industry consensus data.

5. Data related to AI privacy: From IDC, Statista and network security public reports, including "85% of cities around the world deploy AI face recognition, 90% of user data is collected by AI companies, and the accuracy of AI personal behavior prediction is 78%."

6. AI risk probability data: from OpenAI, Anthropic risk assessment reports and Cambridge AI governance survey, including various risk probability statistics such as mass unemployment and deep forgery.