Canva cuts 1,271 roles as studies question the AI job-cut story

Canva cuts a fifth of its workforce

Canva announced job cuts and a sweeping overhaul, Bloomberg reported on 3 November 2026. Chief executive Melanie Perkins said in a note to employees that 1,271 people, or 22% of staff, will lose their jobs, change roles or relocate to central hubs. The note was posted on the company's website. The report said artificial intelligence is forcing a structural reckoning at one of Australia's highest-profile technology companies. Bloomberg described the move as growth slowing for the first time at the design software company.

A smaller workforce will support the design tools Canva's users rely on. The company will consolidate roles into central hubs.

Canva has not said how many affected staff will leave, change roles or relocate. It has not said what the overhaul costs or saves.

The research says AI changes tasks, not whole jobs

An IIT Madras white paper argues that AI automates specific tasks inside jobs. It does not cut entire professions, the paper says. The paper was published on the institute's website on 26 October 2026. The paper's title is "How AI Is Reconfiguring Work in India". It was presented at an AI governance conclave on 22 October. Prof Balaraman Ravindran, who heads the Centre for Responsible AI at IIT Madras, said the work challenges the idea that AI always takes away jobs. "Most jobs have tasks and only some of those can be undertaken by AI," he said. The authors identify seven pathways through which AI can transform work, including cost-driven automation and tasks that need human judgment or regulatory accountability.

Automating routine junior-level tasks could weaken traditional career pathways, the paper warns.

The paper sketches an "AI-cost geography". Cheap, task-specific tools spread through small businesses, while advanced work in legal research and medical imaging stays with large, well-funded organisations. The authors call their framework a hypothesis that needs testing. They recommend affordable AI tools, India-specific research and stronger reskilling routes.

Hiring data shows a gap between buying AI and using it

The iCIMS and Kyle & Co. study surveyed 500 HR and talent leaders and 1,000 job seekers in the US. It found 45% of HR leaders rated their AI implementation highly successful, while organisations put only about half of the AI capabilities they bought into production. Higher-performing organisations used an average of six of seven AI hiring capabilities, against 2.5 among lower performers. About a quarter of respondents had deployed agentic AI across multi-step hiring workflows at full scale. That ran at 42% among higher performers. It ran at 9% among lower ones.

Job seekers see less than employers think. Of those who believed AI had screened them, only 32% said employers disclosed it clearly, and 41% suspected undisclosed use.

Kyle Lagunas is founder of Kyle & Co. He said the gap is not between organisations that have AI and those that do not. "It is between buying capabilities and building the skills, data ownership and operating practices to put them to work," he said.

Workers are learning AI without their employers

A University of Phoenix white paper published on 31 October 2026 draws on its Career Optimism Index to describe how workers pick up AI skills alone. Half of workers report learning to use AI independently, without formal employer training, and 60% want more guidance with the tools. On the employer side, 62% say employees are developing AI skills faster than their organisations can adapt. Another 48% worry about keeping AI-fluent staff.

A separate study examined 45 knowledge workers at a US public university. It found that deliberately asking AI to challenge their ideas improved their performance. Monideepa Tarafdar, a professor of information systems at the University of Massachusetts Amherst, calls the approach generating friction.

Tarafdar said chatbots drift toward agreement, telling users what they already think, and warned that shared tools push thinking toward what she called "collective homogenization". "Don't tell me things I agree with," she said. "Tell me things that I don't agree with or disagree with me or try to poke holes in my argument".