AI comes for your work, much faster than anyone thought

Each week brings another series of layoffs led by AI. In May, Microsoft licensed more than 6,000 Software engineers relied on AI for generation and code development. This same month, IBM cut thousands HR jobs. In February, Meta licensed 3,600 Employees – approximately 5% of his workforce – when they restructured around an AI -ST -ST strategy. These layoffs are not isolated incidents; These are signs of seismic change in the world economy.

Last week, deposits for unemployment benefits tick up Since last fall, companies ranging from Procter & Gamble to Starbucks saying that they provide great layoffs. The share of this is due to the Trump trade war is uncertain, but the rise in automated systems and focused on the work by heart does not help.

Welcome to the immediate disadvantage of the age of AI: economic displacement. And if that looks bad now, consider that we have not reached the so -called general artificial intelligence, the next big phase of the AI ​​era. At this stage, AI can understand, learn and apply knowledge about a wide range of tasks, just like a human. AGE would be capable of reasoning, problem solving and adaptation to new situations in any field without being reprogrammed.

While many experts believe that AC is still in decades, an increasing number of experts say that this is likely to happen in the next five years.

Dario Amodei, CEO of Anthropic, made the headlines last week When he repeated his warnings That the AG level systems could emerge in the two to three years. Daniel Kokotajlo, a former research analyst who left Openai on the grounds that the company did not take security risks quite seriously, said in a report published in May that AGI could arrive at the end of 2027.

And Ray Kurzweil, futuristic and engineering director at Google, Continue to predict act Will be reached by 2029, a date on which he reaffirmed last year in “the singularity is closer”.

“In my opinion, we are roughly on the right way of Act at the level of man by 2029,” said Ben Goertzel, CEO of Singularitynet, a decentralized and open source platform that allows AI agents to cooperate, share data and offer services on a blockchain-based network.

And after that? Goertzel and others believe The jump to “super intelligence” could only be a question of a few years: “I think it will only be a few years from an act of a human level to a super act, because this act of human level will be able to program and invent new fleas and to invent new forms of networking.”

Many experts believe that the final evolution of super intelligence could lead to a deep structural collapse of traditional employment, moving everyone, from C-Suit to doctors, lawyers, scientists and doctoral researchers, and even very entrepreneurs who are currently building their fortune on AI.

But the passage to AG will be sufficiently heartbreaking.

We are far from ready

The general artificial intelligence will extend far beyond the automation of routine tasks, with the ability to reason, adapt and surpass humans in almost all areas.

“Once AI becomes slightly smarter than humans, we will see massive unemployment,” said Goertzel Decipher. “It can start with junior white -collar jobs, but I think it will quickly extend to plumbers, electricians, concierges – everyone.”

Goertzel notes that AI has outperformed doctors in diagnostic precision for years, but industries and health care have resisted change due to institutional power and license requirements.

“Entry-level jobs have no one defending them,” he said. “Elderly people in powerful positions can protect their roles – and they are controlling how AI is deployed. So, of course, they will not replace themselves with AI. ”

According to Goertzel, the AI ​​has not disrupted blue collar work as aggressively as white -collar work because the physical equipment has still catch up. Because AI in software has largely exceeded robotics, this disparity helps to explain why the white collar jobs have brought the weight of dismissals based on AI, while the roles in blue collar remain relatively intact, for the moment.

Half of the input white collar work could disappear

In a recent interview, the CEO of Anthropic, Amodei, warned that the disruption of the employment of AI is not in the decades – it already occurs and will accelerate quickly.

He estimates that up to 50% of the roles of entry -level white collar could disappear within one to five years. These roles include positions at the start of their career in law, finance, consultation, marketing and technology – tasks which once offered stable ramps in professional careers.

As AI tools, the tools are increasingly managing analysis, writing, planning and decision -making, many of these human positions are made obsolete. In a separate interview with CnnAmodei reiterated his complaint, warning that change would occur earlier that humanity cannot be prepared.

“What strikes me in this AI boom is that it is larger, it is wider and that it evolves faster than before,” said Amodei. “Compared to previous changes of technology, I am a little more worried about the impact of work, simply because it happens so quickly than, yes, people will adapt, but they may not adapt quickly.”

White collar jobs already threatened

If you work behind a screen, you are already in the Blast AI department.

“The most exposed jobs are those that require higher education, paying more and involving cognitive tasks,” said Tobias Sytsma, Rand Corporation economist, said Decipher. “Historically, this type of exposure to AI has been correlated with employment reductions.”

According to An April 2025 report By the Federal Reserve Bank of New York, graduates in computer engineering are faced with double the unemployment rate of majors in art history at 3% against 7.5%, respectively.

Here are some of the jobs which, according to economists, are the most immediately exposed to AI:

• Software engineers: Companies use AI to generate, examine and optimize the code. In May, Microsoft improved its Github Copilot to a full AI agent.
• Human resources: AI is used to filter curriculum vitae, assess employee performance and write termination mail.
• Parajurists and legal assistants: AI can summarize case law, examine contracts and write results.
• Customer service representatives: Chatbots are used to interact with customers and manage routine support tickets. With voice And AI Video Becoming widely available, call centers are being removed.
• Financial analysts: AI models can analyze Massive amounts of data and generate reports more effectively and precisely than humans.
• Content creators: Writers, publishers and graphic designers are already in competition with generative AI tools. In 2023, the writers Guild of America were on strike, the protections of the AI ​​being a key problem.

“Our research shows that these are mainly white collar jobs – those that require higher education, paying more and involving cognitive tasks – which are the most exposed,” said Sytsma.

However, health professionals are relatively protected due to the regulations. “Health care seems to be the one where these workers inside the moment. However, exposure to these tools increases. What is happening then remains uncertain. ”

“We could automate most jobs without achieving AC full at the human level, because most of the work is repetitive and based on previous examples – and has managed well,” said Goertzel. “Jobs requiring large imaginative jumps are more difficult to replace, but most economic activities are not based on this.”

Goertzel suggested that superintendent could even automate and replace political leaders. “Even the presidency could be automated,” he said, “but political standards make these limits-for now.”

Whether AI brings us to a post-work utopia or a deeply unequal dystopia, one thing is clear: if you always think that AI arrives first for the blue passes, then you are already behind the curve. He comes for the office next to you. And that doesn’t wait.

Edited by Andrew Hayward