Nintendo adopts wait-and-see on generative AI | Opinion

Every creative industry is currently grappling with the question of what generative AI means for them. The step-change we have seen in the capacity of computer algorithms to generate imagery, audio, and text is a significant one in many ways.

Philosophically, it pushes back the boundaries of what computers can do into an area that makes many people uncomfortable, because this kind of "creativity" has so often been considered an inherent part of what separates humans from animals, let alone computers.

In more substantive terms, it opens up possibilities for productivity while also raising serious questions about the commercial sustainability of creative professions and businesses. Battle lines are being drawn, but a lot of people aren't sure which mast they want to pin their colours too, holding both hopes and fears for this new technology.

The games industry, thus far, has been relatively cautious about generative AI – there have been a fair few optimistic statements about the technology, but not a lot of cases where it seems to have been deployed to any particularly significant effect in a development process.

Major companies are taking a wait-and-see approach to some degree, and most of the noise about the use of generative AI (both in terms of eager adoption, and angry backlash) has been concentrated around unknown indie upstarts.

Nintendo, unusually, has become one of the first companies to actually come out and take a stance on generative AI – making comments in an investor Q&A to the effect that it doesn't have any intention to use generative AI in its development processes for now. Nintendo is careful to specify that it's talking about generative AI, noting that games companies have been at the forefront of using other kinds of AI in their products for years.

Its decision on generative AI isn't coming from some kind of luddite refusal of new technology – I have no doubt that there's a ton of work going on with other kinds of AI at Nintendo, especially the more proven technologies that now seem to be relegated to being called "machine learning" or something similar since LLMs and other generative technologies took over the "AI" label. (None of these things are actually AI in any strict sense of the word, but that linguistic battle was lost quite some time ago, I fear.)

So why is Nintendo – usually so guarded and careful to avoid taking a strong stance on anything – so willing to eschew generative AI in this way? There are, I think, a few reasons, and they're all worth thinking about for other companies that are wondering about the position they should take on this technology – many of them balancing once-bitten, twice-shy sentiments in the wake of the implosion of hyped technologies like blockchain and NFTs against serious FOMO around the rapid technological progress of systems like ChatGPT.

First and foremost, there is a value statement to the position Nintendo is taking here, and as brief and throwaway as the comments may seem, the words were certainly chosen to reflect a value that Nintendo wants to project to its investors and to the world.

To wit; the company fundamentally doesn't think that you can create the kind of valuable IP that is the bread and butter of its long-term business by outsourcing any significant part of your creative process to an algorithm. That's not an unreasonable belief, nor a luddite one; it is a statement of where Nintendo thinks its value as a company lies, and how it plans to build and defend that value in future.

Nintendo is one of the most hard-nosed companies in the business, but it's also a company where top creatives hold significant managerial power – far more so than at most publishers. It is also a very old company with a very long-term perspective.

It's common to see this aspect of Japanese companies romanticised as a contrast to the supposedly more short-term thinking of Western firms, a narrative that veers into misty-eyed Orientalism all too often, but there is at least a kernel of truth to the idea that the executives of a company like Nintendo feel a responsibility to safeguard the company's long-term future even if it's at the risk of a few percentage points of performance in this quarter or the next.

Even if adopting a new technology could deliver a string of great quarters for Nintendo, if it creates a longer-term risk to the company's value, there's a strong lobby within the company for holding back. This was seen very clearly with smartphone games, which it engaged with slowly and hesitantly – much to activist investors' fury – and which it largely pulled back from once it became clear that the Switch was a major hit.

Nintendo will likely resist any similar pressure to engage with generative AI if it has any fear that doing so could damage its value in the longer term. At the same time, though, we should recognise that this is a pretty easy position for Nintendo, specifically, to take.

It's a Japanese company and most of its development studios are located in Japan; as Alicia Haddick's article this week pointed out, layoffs are much harder for Japanese companies, both legally and culturally.

Nintendo fundamentally doesn't think that you can create the kind of valuable IP that is the bread and butter of its long-term business by outsourcing any significant part of the creative process to an algorithm

This defuses one of the quietly understood arguments for generative AI – the notion that a company could cut costs by deploying AI tools to do work that had previously been done by junior staff. Of course, advocates for generative AI in the industry strenuously deny that there's any intention of replacing staff with these tools – and for now, that's an easy argument to make, because anyone who's tried to integrate them into a creative workflow can tell you that AI tools aren't remotely good enough to replace skilled workers right now.

However, it's also perfectly obvious that plenty of business leaders are eyeballing the prospect that in a few years' time they'll be able to cut development costs and timescales by automating away various job roles.

In Japan, that doesn't really have the same allure – and it's not just because it's harder to fire workers, it's also because many senior people in Japanese game companies view working up through those junior roles to be a core part of the staff development process. If AI is doing the jobs of junior designers, coders, or artists, then where does the next generation of experienced people to fill senior roles come from?

There's one more aspect to Nintendo's thanks-but-no-thanks response to generative AI which is made very clear in its statement – the company is concerned about the copyright and IP uncertainty around anything created with generative AI.

This is a problem that's brought up every now and then with regard to generative AI, and is usually hand-waved away by evangelists, but it's a very real concern – especially if you're a company whose entire value is based on a foundation of self-created and self-owned IP.

Right now, there's very little legal clarity about whether material created by generative AI can be copyrighted – to say nothing of the very thorny questions around the rights to the training data that have been used to actually create most generative AI models, and what legalities arise when those models spit out something that closely resembles data from that training corpus.

The whole thing is a minefield, and the adoption of the technology is moving much faster than the law – which needs to be clarified both at the level of individual countries and at the level of international copyright conventions. This almost certainly means that some people (on one side or another of the situation) are going to get dealt a very tough hand when the law finally does catch up. Nintendo has no interest in being on any side of that situation; the only way to be assured of winning is not to play.

It's obvious that plenty of business leaders are eyeballing the prospect that they'll be able to cut dev costs by automating away various job roles

Of course, Nintendo isn't the Japanese games industry's appointed spokesperson, and other companies will make their own decisions – I can't imagine Square Enix, which jumped around every blockchain and NFT flash in the pan like a ten year old hopped up on sugary drinks and set loose in a toy store, making the boringly mature choice to leave generative AI alone until its issues are clarified and sorted out.

However, Nintendo's stance isn't isolated either. It won't be surprising to see other companies in Japan and elsewhere making measured, but quite firm, decisions to leave this technology be, at least until the issues around it are clarified and the question of how to fit it into creative processes and pipelines is figured out.

Generative AI is both a promising and a troubling invention in various different ways, but it's likely an inevitability that it will become part of the landscape of creative industries in the end – like any other technology which has revolutionised creativity, no amount of protest or misgivings changes the fact that it exists now.

Toothpaste can't be squeezed back into the tube. For companies whose unique value is entirely built upon their ability to create and effectively steward much-loved IP, such technological disruptions to the creative process have to be considered carefully – and Nintendo, at least, appears to have decided that creativity is one part of its business that isn't in need of any kind of Silicon Valley disruption.