Anthropic boss and co-founder Dario Amodei has warned that the rapid advance of AI could upend the workforce “faster and more broadly than anything we’ve seen before”, even as his company embraces the technology to transform how it operates.
Speaking on stage at Salesforce’s Dreamforce conference in San Francisco, Amodei announced that AI tools were now deeply embedded in Anthropic’s workflows, reshaping how teams operate.
“I made this prediction that, you know, in six months, 90 per cent of code would be written by AI models”, he said. “Some people think that prediction is wrong, but within Anthropic and within a number of companies that we work with, that is absolutely true now”.
He said on Thursday that he was sceptical about people’s ability to adapt to the adoption of AI, with “jobs changing so fast”.
Despite his caution, Amodei said he remains optimistic about what’s ahead.
“I’m one of the loudest in voicing concerns about safety”, he claims, “but I’m also an incredible believer in the upside – especially for curing disease and advancing science. We just have to make sure we get there responsibly.”
The workforce challenge
Amodei stressed that automation does not mean fewer humans, but rather, different kinds of work.
“If Claude is writing 90 per cent of the code, what that means, usually, is you need just as many software engineers, you might need more, because they can then be more leveraged”, he claimed.
“They then can focus on the 10 per cent that’s hardest, or supervise a group of AI models. It’s a rebalancing, not a replacement”.
Still, he acknowledged rising anxiety across the economy. “Jobs are changing so fast”, he repeated during the keynote. “I do worry that across the economy, we’ll have a kind of labour disruption that operates faster and is broader than anything we’ve seen before”.
Amodei’s remarks follow his warning at the Axios AI+ DC Summit last month, where he said that AI could displace “up to half of white-collar jobs” within five years.
As AI adoption grows, so do concerns
Anthropic’s own ‘economic index’, published in September, found that 77 per cent of businesses using its Claude AI assistant are now deploying it for automation, rather than for collaborative uses like research or analysis.
That trend aligns with a broader pattern across the tech sector, where, despite near-universal AI adoption, most companies are still struggling to achieve broad productivity gains.
A Stanford study this summer found that employment for software developers aged 22 to 25 has fallen almost 20 per cent since late 2022, reflecting how AI tools are absorbing routine coding and data tasks once assigned to junior staff.
“AI is doing a lot of what entry-level workers used to do”,said Harvard Business School professor Christopher Stanton, warning that wages could fall as employers reimagine early-career roles as supervisory rather than creative.
At Dreamforce, Amodei framed the moment as both opportunity and warning. “Comparative advantage is powerful”, he claimed. “I’m optimistic about complementarity in the short run. But as we get out to two to five years, we’ll see enough disruption that it could shake up the economy”.
Founded in 2021 by Dario Amodei, among other OpenAI spinouts, Anthropic has positioned itself as one of the world’s most prominent AI research companies.
Its Claude AI model now powers features within Slack and across AWS cloud products.