Browsing Reddit today, I came across an explosive hot post.
That's right, Anthropic's CEO Dario Amodei, also known as the "Father of Claude," has made another striking statement:
Artificial intelligence could eliminate half of junior white-collar jobs within the next one to five years—leading to an unemployment rate surge of 10%-20%.
He stated bluntly that AI companies and governments must stop "sugarcoating" and face reality: various white-collar industries, including technology, finance, law, and consulting, especially entry-level positions for young people, are being pushed to the brink by AI.
Following this image, the editor tracked down the original source—an in-depth report from the well-known American media outlet Axios, with an equally chilling title—"Behind the Scenes: The AI-Induced White-Collar Massacre."
After reading the full article, it became clear: this isn't Amodei being dramatic; it's likely a view he has repeatedly emphasized in private, and it's becoming a growing consensus within the tech community.
Axios reporters accurately identified the "elephant in the room": everyone is ignoring the unemployment problem AI will bring—legislators are unwilling to face it, the U.S. government avoids discussing it, and AI companies are afraid of alarming people.
Axios had to "persuade him for a long time" before Amodei agreed to speak publicly. So, what did Amodei say?
Claude 4 approaches human-level capabilities: it can bring benefits, but also 20% unemployment.
Amodei is contradictory.
At the recently concluded Claude conference, as the "Father of Claude," he personally released the new version, Claude Opus 4—its programming capabilities are considered to be approaching human level. He is proud of the technological breakthrough but does not shy away from the potential costs:
This technology could bring about cancer cures, 10% annual GDP growth, and balanced government budgets—but it could also put 20% of the population out of work.
He warned:
"Most people simply don't realize this is about to happen," he said. "It sounds crazy, people just don't believe it."
Amodei's core point is: stop treating AI with a "toy mentality." He said, don't just scratch the surface using AI; once you truly dive deep, you'll find it can already handle the work of researchers, analysts, and a large number of entry-level white-collar positions.
His wording was very severe—according to Anthropic's preliminary assessment, half of entry-level white-collar jobs could be eliminated within five years, and the unemployment rate could soar to 10% to 20%.
And these words are not from an alarmist critic, but from someone who is personally building this technology. When reporters asked, "You're building this thing while also warning about the destruction it might cause; what are your thoughts internally?"
He replied: "I have a responsibility to develop this technology, and I also have a responsibility to push governments, legislators, and other AI companies to think about—how to prepare American workers and how to protect them."
But Amodei also understands the absurdity of reality:
"We say 'everyone should be worried about what we're building,' but people's response is: 'You're just hyping it up.'"
So he retorted: "What if we're telling the truth?"
The "Bloody Scenario" Amodei describes: Collapse overnight
As compiled by Axios, Amodei's unemployment warning has been reduced to a clear chain—not a gradual occurrence, but a sudden collapse:
1. AI giants like OpenAI, Google, and Anthropic are continuously enhancing the capabilities of large language models (LLMs), which can already handle more and more human tasks.
2. The U.S. government fears falling behind China or is unwilling to disturb the labor market, thus remaining silent, neither regulating nor publicly warning.
3. The American public lacks awareness of AI's threats and is not concerned.
4. Then, almost overnight, companies find that replacing human labor with AI saves significant costs, and begin to fully replace human employees.
By the time the public realizes it, it's too late.
Agents "at the table," interns out
Meta CEO Mark Zuckerberg publicly stated on Joe Rogan's show earlier this year:
"By 2025, we at Meta might have an AI that can handle the work of a mid-level programmer."
Shortly after, Meta announced 5% layoffs.
OpenAI's Sam Altman also said that this year is the year of agents, and agents are joining the workforce.
In fact, hundreds of tech companies globally are racing to develop "agents"—these AIs can perform tasks continuously, stably, and at low cost, essentially acting as "virtual interns" who can work overtime anytime and require no training.
However, this wave of replacement primarily affects entry-level positions.
A few days ago, we reported on a case: a senior engineer successfully debugged a four-year-old legacy bug using Claude, but he admitted:
"Claude is, at most, a junior engineer in terms of programming. If you asked me to choose between AI and 30 senior engineers for a team partner, I believe no one would choose AI."
It sounds like "humans have not been completely replaced," but for young people just entering the workforce, this reality remains harsh.
The Axios report quoted former White House advisor Steve Bannon's comment:
"No one is seriously considering the reality that administrative, managerial, and technical jobs for people under 30 will be destroyed," Bannon said. "These are the most critical training positions in your 20s."
How to "turn the wheel"? Amodei proposes four suggestions
Amodei believes that the overall trend is irreversible, but there are still remedial measures.
"You can't just stand in front of the train and try to stop it; the only way that works is to turn the steering wheel—to deviate the train by 10 degrees from its original direction. This can be done, but we must do it now."
His four suggestions include:
1. Increase public awareness and provide early warnings
2. Slow down the rate at which jobs are replaced
3. Enhance AI literacy among government officials
4. Initiate policy imagination for a "post-AI society"
When discussing the fourth point, taxation, Amodei also put forward a bold and unpopular proposal:
"Token Tax" plan:
Whenever users invoke AI models and companies profit from it, approximately 3% of the revenue would be collected as an "AI dividend tax," managed centrally by the government, and then returned to society through some mechanism.
He admitted:
"This is obviously not in my economic interest, but I think it's a reasonable way to solve the problem."
If AI's growth rate is truly as he predicted, the scale of this 'AI dividend' will be in the trillions of dollars.
Concluding thoughts
On Reddit, some netizens praised Amodei's "radical statement" as a rare display of courage—few AI company CEOs dare to openly acknowledge the impact they are creating.
They do say "workers will be replaced," but they also emphasize that it won't cause chaos. For example, Altman says, "Yes, many jobs will be replaced," but "overall it's good for humanity." If you publicly state that this will lead to mass unemployment, it somewhat damages your public image.
On YouTube, this interview video has also garnered thousands of comments. The most upvoted one reads:
This sends chills down my spine. "Half of white-collar entry-level jobs gone in five years"? That's not the future, it's almost tomorrow. I've been in HR for ten years and can already feel this shift. The scariest part? Most people are still sleepwalking into this crisis.
Amodei's statement may not stop anything, but it successfully generated discussion—allowing the public to confront a problem that previously only existed in science fiction movies.
And for everyone involved, in the age of AI, how we plan our career growth and life path may be a more realistic and urgent issue.
If AI will restructure the entire world, does individual "long-termism" still hold true?
Do you think Amodei's prediction is an exaggeration? Feel free to share your thoughts in the comments section.
References:
1.https://www.youtube.com/watch?v=OKCD2dmcjsQ
2.https://www.reddit.com/r/OpenAI/comments/1kxnyun/dario_amodei_says_stop_sugarcoating_whats_coming/