Artificial intelligence blame for video-game industry layoffs may be misguided
Amid a wave of high-profile layoffs in the video-game industry , including a recent studio shutdown at Ubisoft Entertainment SA’s Halifax office, artificial intelligence is emerging as a lightning rod for debate because some are linking its use to job cuts, but others say the decisions are being driven by dollars, not technology.
Multiple researchers and industry insiders say generative AI is not capable of wholesale workforce replacement in the industry, so the business decisions driving layoffs may have little to do with AI at all.
Derek Nowrouzezahrai, a McGill University professor and chair of the Canadian Institute for Advanced Research in AI, said the business case for widespread job losses due to AI is being overstated. The technology can speed up certain tasks such as interface design and help with finding bugs, but he said it lacks the control, context and reliability needed to replace skilled workers in real production environments.
“These tools are extremely powerful,” he said. “But there’s only so far you can get without careful and expert human supervision, guidance, and interaction.”
Nowrouzezahrai said developing modern games requires precise iterative changes that must remain consistent across large interconnected projects, something current AI systems struggle to handle.
“If I say, ‘Move it two centimetres to the left,’ that’s a very specific change. That level of precision and control is still difficult for these systems,” he said.
Nevertheless, Montreal-based studios, including Eidos Interactive Corp. and Keywords Studios PLC, have shed nearly 500 positions over the past 18 months. And Hollywood’s video-game performers went on strike last year over AI and pay issues while studios shut down and more than 10,000 people lost their jobs.
Nowrouzezahrai said the core limitation of generative AI is not its ability to produce content, but its inability to understand the broader creative and organizational context in which that content is used.
“When an art director says, ‘I need this box to look more like the first project,’ there’s years of shared context behind that instruction,” he said, referring to the long-standing working relationships, artistic conventions and prior decisions that guide human collaborators, but are invisible to AI systems.
From a business perspective, he said companies risk misjudging both the capabilities and the failure modes of the technology.
“These systems can fail catastrophically and with great confidence,” he said, adding that experienced workers are still needed to supervise, test and correct AI-generated output. “For many high-value tasks, a human in the loop remains essential.”
Nevertheless, AI investment in gaming is big business. It is projected to grow at a compound annual growth rate of more than 36 per cent and to reach US$1.65 billion by 2033, according to a Cognitive Market Research report.
In 2025, the market value of AI in Canadian games was estimated at approximately US$138 million, up from roughly US$40 million in 2021.
Despite that level of investment, one in 10 game developers was laid off over the past year, according to a 2025 report by the Game Developers Conference. It also said 30 per cent of developers believe AI has a negative impact on the video-game industry, a 12-point increase compared to the previous year.
Nowrouzezahrai is also concerned about how AI is influencing workforce decisions, particularly at the entry level. He said eliminating junior roles in the name of automation could undermine the industry’s long-term sustainability.
“The only way people build expertise is by doing those simpler early career tasks,” he said. “If you remove that step, you’re cutting off the pipeline that produces senior talent.”
Earlier this month, Jaana Dogan, principal engineer on Google LLC’s Gemini API team, sent shockwaves through the technology industry with a viral post on X claiming that Anthropic’s AI coding tool, Claude Code, had generated a system in one hour that her team had spent nearly a year developing.
But Nowrouzezahrai said it’s not all about speed.
“This is not a new story. We went from slide rules to calculators, from calculators to spreadsheets, and humans were still in the loop every time,” he said. “The human still has to understand the tool, direct it and take responsibility for what it produces.”
This past summer, a Google Cloud survey of 615 game developers across five countries said 87 per cent are using AI agents to streamline repetitive tasks, with 44 per cent applying them to optimize content and decision-making in text, audio, code and visuals.
Developers said AI frees them up for more creative work and could reduce long-term development costs, but it also raises concerns about job losses, intellectual property and workforce planning.
Others say the story is less about technical limits than it is about corporate behaviour.
Matthew Guzdial, a computing professor and Canadian Institute for Advanced Research AI chair at the Alberta Machine Intelligence Institute at the University of Alberta, leads a team of about 15 students exploring AI-driven creativity in games. He studies procedural content generation, a field that has been experimenting with algorithmically generated game content for decades.
He said AI is primarily being used to support developers rather than replace them and is often mischaracterized as the cause of layoffs.
“AI is certainly involved, but if you look at the claims, AI replacing workers doesn’t seem to be actually shaking out. It’s being used as an excuse to cut headcount,” he said. “A bunch of people who are not actually experts on these technologies are trying to convince people they can do more than they can.”
Guzdial said AI tools rarely replace the iterative, complex work of developers.
“Producing code is easy. Debugging code is hard,” he said. “Similarly, artists and musicians don’t just create one image or track; they iterate constantly within technical and design constraints.”
He also pointed to the pandemic-driven hiring boom, which inflated head counts in the industry and created unsustainable expectations.
“The factor rate increased about 25 per cent during COVID-19 and now we’re seeing a reduction of about 25 per cent,” he said. “It’s less about AI, more about overhiring and studios shuttering.”
Guzdial said entry-level jobs such as 2D artists and writers are the most exposed to AI-driven automation while more complex tasks such as debugging, 3D asset creation, game design and narrative remain secure.
“It’s not AI that’s killing jobs,” he said. “It is the profit motive that’s killing jobs.”
Guzdial also predicts a market correction in AI investment, calling the current hype an unsustainable bubble.
“There is real value to these technologies right now … but it will be very much decreased,” he said.
Alex Kearney, co-founder of Edmonton-based Artificial Agency Inc. and a former DeepMind Technologies Ltd. researcher, said the technology is creating new opportunities even as it stokes fears.
“The games industry is a strange beast,” she said. “People are looking for easy explanations for layoffs and, right now, AI is an effective boogeyman. It’s changed the way that I program, but it hasn’t reduced the work that I’m doing or the demand for programmers.”
For Guzdial, Nowrouzezahrai and Kearney, the message is clear: AI in games is not a job killer on its own.
“Investors and executives have been sold the beautiful dream of Silicon Valley,” Guzdial said. “There’s a bubble; this isn’t the first AI summer and it won’t be the last.”
• Email: arankin@postmedia.com