The videogames industry, like so many others, moves in cycles – at least, so the conventional wisdom goes. A few tough years should be followed by a boom, and vice versa; growth leads into consolidation leads into growth.
In terms of hiring and layoffs, the industry has certainly experienced a few tough years. Many thousands of jobs were lost in wave after wave of layoffs and studio closures. Some companies were growing through those years, but not nearly enough to absorb all the staff laid off. The industry overall shrank its headcount significantly.
You might expect that the turn of the cycle would see companies starting to expand again and building out new teams by this point. If so, the employment survey released this week by InGame Job makes for disappointing reading, but largely only confirms what’s already clear from a glance around the industry: the green shoots of new growth are few and far between.
While there will always be reasonable questions to be asked about the representativeness of a voluntary survey (not least that people who have experienced disruption in their careers may feel more inclined to respond to surveys about employment, which can introduce systematic biases in the data), the year-on-year comparisons the survey offers are less susceptible to such issues, and very instructive.
They paint a picture of an industry in which, compared with last year, more people are involuntarily unemployed, and fewer people are voluntarily changing jobs. That latter data point, showing that people are choosing job security over career progression, is a clear indicator of a climate of insecurity.
For those who are forced to look for work, it has become a hirer’s market. Despite the long-standing concerns about skill shortages in the industry, the reality (in European countries at least) is that there are so few open positions and so many applicants that employers have been able to push down salary offers for new hires, and freeze wage increases for existing staff.
So why is this happening? Why hasn’t the turn of the wheel brought us back into a growth phase of the cycle?
First, we have to look to the reasons why the industry hit such a catastrophic wave of layoffs and closures in the first place. The causes here are primarily economic, unsurprisingly, and the trigger was exogeneous – the sharp rise in borrowing costs after many years of interest rates hovering around zero.
Higher interest rates dramatically curb companies’ appetite for risk-taking, and few industries were quite as vulnerable to a decline in risk appetite as the games business. This is due to the industry’s own more long-term economic problems, namely rising development costs despite largely stagnant per-user revenues on most titles. For years, strong audience growth threw a rug over that problem, but now audience growth has stagnated too.
The result is an industry that’s not just hit-driven – that was always the case – but one in which titles that aren’t hits struggle to make any money at all. When risk appetite was high and borrowing was cheap, lots of companies were happy to throw money at projects and studios in the hope of generating the next Fortnite or the next GTA, and the financial failure of most projects was an accepted cost. When interest rates were hiked and risk profiles shifted, the industry’s unstable fundamentals were left terribly exposed.
So to the question of why the cycle hasn’t turned upwards, we can consider the fact that this basic situation hasn’t actually changed in the past few years. If layoffs have slowed down, it’s because most companies have already pared back to the bone – not because the conditions that created those layoffs have changed. Interest rates remain high. Risk appetite remains dulled.
There’s another factor in the mix now as well, though. It’s an open secret that generative AI tech is being experimented with widely around the industry, but attitudes towards its potential vary massively between developers on the creative coalfaces and people in executive and managerial roles.
It’s rare to meet a developer who hasn’t played with AI tools and tried them out in various parts of their workflow – the games industry isn’t noted for its luddites, after all – but their assessment of them is almost universally cautious to the point of being lukewarm. This isn’t because they’re afraid AI will steal their jobs. Their concerns are more practical; they can see how error-prone AI is even in the tasks to which it’s best suited, such as code generation. In any field, the output from AI tools rapidly gets worse and mistakes become more common as the task at hand becomes more specialised; a huge problem in a field that’s inherently specialised, like game programming.
Developers’ concerns about AI adoption tend to focus on those errors and on the very real likelihood that large-scale, incautious use of AI is just going to build up a massive workload of broken code and assets that will ultimately need to be fixed by humans – potentially a slower and more expensive process than just doing it by hand in the first place. Promised productivity benefits for professional creators and developers using AI as a “copilot”, meanwhile, simply fail to materialise in properly controlled studies.
Unfortunately, that caution is not universally shared at higher levels in the industry – where many have been caught up in the infectious optimism of sales pitches from AI companies currently burning through investment money at unprecedented levels (all that cash that’s pointedly not being put behind risky game projects still has to go somewhere, after all).
AI companies are motivated solely by juicing their numbers to unlock the next tranche of venture capital investment and have made outrageous claims in service of that goal. They paint a picture of a future almost within reach where experts’ productivity will be boosted by orders of magnitude by AI assistants, while all but the most complex of tasks won’t even need experts at all. The gap between idea and implementation will be sewn shut by the tread of all-knowing AI.
AI tools will, of course, progress and improve from where they stand today – but the gap between what’s actually possible now (marginal usefulness in specific scenarios with careful expert supervision) and what’s promised in the near future is incredible, in the most literal of senses.
Yet this connects back to the industry’s dire employment circumstances, because on multiple occasions in recent months I have heard – directly or second-hand – of senior executives pushing back against hiring more staff on the basis that they wish to “wait and see how AI turns out”.
This is not unique to the games industry by any means – it’s happening across almost every industry that involves people sitting in front of a computer in any context, and having an even bigger impact in some other fields than in games. It arguably hurts more here, though, since it comes in the wake of those years of massive layoffs; every company holding off on hiring staff on the off-chance that ChatGPT will be able to do their job by next year is another skilled worker sitting unemployed, re-evaluating their career choices, and potentially lost to the industry for good.
I don’t think AI can be blamed for the bulk of the current employment climate – economic factors like interest rates and fears of spending-led recession are much more directly responsible. However, the dream of AI-driven productivity leaps is definitely depressing the appetite for hiring even further in many companies.
The effects of this may be felt for years to come. Games are developed on cycles that now run for upwards of half a decade in many cases, which means that studios which place their bets wrongly at this point in the cycle – such as those wagering too heavily on AI revolutionising development in ways that current evidence simply doesn’t support – are likely to be paying for those mistakes for a long time.
After the peak years of the COVID-19 pandemic, we saw a delayed impact on game release schedules. Studios that had been able to adapt quickly to things like remote working were able to release well-polished games into a surprisingly clear release timeframe because other titles had faced massive delays. Some of the biggest commercial hits of that era were reaping precisely those rewards.
Similarly, the bets placed now – on generative AI, and on industry growth more generally – will pay off not today or tomorrow, but in three to four years’ time. It seems likely that healthy caution on the prospects for AI in development will pay significant commercial dividends at that point. There’s no better way to get started on that than crafting hiring policies to take advantage of the huge surplus of talent on the job market right now.
