Anthropic Researchers Warn of ‘Terrible Decade’ Ahead Over AI Threat to Wipe Out White-Collar Jobs

Artificial intelligence accelerates faster than society cannot adapt, and anthropic researchers warn that this could trigger a turbulent decade, in particular for white passes whose work is increasingly risky to be automated.

The IA company, which has recently published some of the most advanced large language models to date, has said that current technology is already able to replace numerous entry -level office roles and that the situation could get worse in the coming years.

Sholto Douglas, one of the best researchers in the company, told the IA Dwarkesh Patel commentator that a drop in the employment of white passes was not only possible – it is probably inevitable. He said that this change could start within two years and become impossible to ignore in the five.

“There are all this spectrum of future crazy people,” said Douglas. “But the one I think we are almost assured of obtaining – is a solid statement to make – is one where, at the very least, you get a drop in white passes at some point in the next five years.”

He added that even if the progress of AI should slow down, the existing models trained on the data specific to the field are already capable of automating a large part of administrative, customer service and analysis functions.

“The current suite of algorithms is sufficient to automate the work of white collars provided that it has enough good data types,” he added.

Dario Amodei of Anthropic had already echoed similar concerns. In a recent interview with CNN with Anderson Cooper, the company’s CEO warned that up to half of all entry -level white collar jobs could disappear within five years due to AI automation. These roles include positions in customer support, parajuristic work, accounting and marketing – jobs formerly considered to be sheltered from mechanization.

Trenton Brcken, another member of anthropogenic technical staff, added that companies should already prepare for large -scale automation of office work.

“We have to expect to see them automated over the next five years,” he said.

Evidence of disturbance already surfaces in the business world. Companies like Shopify and Duolingo have started to reduce hiring for functions that AI now manages, such as content creation, basic coding and support for support tickets. Meanwhile, Klarna, the Swedish Fintech company, has briefly adopted AI to replace more than 700 customer service employees, but had to resume part of this adoption due to performance and confidence problems.

Revalio Labs, an analysis company of the workforce, reported an abrupt drop in job offers for high exposure white collar roles such as data entry, computer support and financial analysis – positions that generative AI models can now perform more and more with minimum human supervision. Companies that seek to reduce costs have already started to turn to tools such as Openai GPT-4, Google Gemini and Anthropic Claude series to automate documentation, data processing and even aspects of software engineering.

The CEO of Nvidia, Jensen Huang, recently added his voice to the growing choir warning employment disturbances. Huang, whose company produces the GPUs feeding most AI systems, noted that “each work” will be affected by AI. He predicted that if some employees may not be completely replaced, they could always find themselves deciding by more productive colleagues taking advantage of AI tools to outdo work flows only.

The new anthropic models, Claude Opus 4 and Claude Sonnet 4 are already showing signs of this future. Early tests show that these systems can write software code, produce legal drafts, summarize long reports and even solve technical problems with a minimum of contribution. A tester said OPUS 4 “was independently coded for almost seven hours” on a complex task, offering an overview of how AI could even move highly qualified knowledge workers.

The wave of imminent automation has raised deeper societal concerns. Unlike the waves of previous disturbances which mainly affected workers in the factory or in the field, the latter targets the middle class – that with university degrees and office careers. Economists say that this change could arouse mass underemployment and wage stagnation, especially in countries where the services sector dominates.

At the same time, the cost savings that AI offer are difficult to ignore for businesses. Replacing an entry -level analyst team with a single AI model has refined on internal data can reduce operational expenditure, increase productivity and rationalize decision -making. But social benefits of mass job losses can force governments to intervene.

Calls for stronger work protections, AI taxation and universal basic income have resurfaced while political decision -makers attack the way of managing the transition. Some experts urge the creation of recycling programs and education reforms aimed at equipping workers with robotics skills, automatic learning and man -centered design – roles always out of reach for today’s AI.

The calendar presented by researchers like Douglas and Amodei already checks. As they warn, the transition could be brutal before it improves. If robotics, which is lagging behind software, does not follow the pace, displaced workers can find themselves incapable of pivoting new roles that require physical work or technical expertise. In this gap, the world can experience what Douglas has called a “fairly terrible decade” – a painful reality where jobs disappear more quickly than new opportunities emerging.

“Imagine a world where people have lost their jobs, and you don’t have new biological research yet. This means that people’s quality of life is not considerably better,” he said. “A decade or two after, the world is fantastic. Robotics are resolved and you arrive at radical abundance. ”