投资者与发行商视角：AI复制品泛滥下的游戏发现危机与新平台机遇

GamesIndustry.biz attended the AI and Games Conference in London this week, where a publisher and investor panel discussed the problems and opportunities of AI when it comes to marketing, funding, and publishing games (the full talk will be available via the AI and Games Conference YouTube channel at a later date).

The panel was made up of James Barnes (global PR lead at Raw Fury), Stefan Metaxa (commercial director at Spiderling Studios), and Lirui Ding (principal at Transcend Fund), and was chaired by Cassia Curran (founder of Curran Games Agency). Some of the key insights from the panel are below, including the industry reaction to AI copycat titles, the growing crisis around discoverability, the rise of new platforms such as web games, and the thorny issue of players' reaction to generative AI use.

AI copycats

The issue of games being rapidly copied using AI has been in the news a lot recently, as more advanced AI models make it easier and easier to copy games. Needle in a Haystack Simulator, for example, was cloned multiple times after the developer released a work in progress video of the concept.

"It's a wild situation, and I completely understand that a lot of small developers who have unique ideas might want to back away from marketing and being too outright with these ideas, in case there are copycats that are lurking and might beat them to the punch," said Barnes. "But I think there's a lot of ways that you can make marketing work for you. You can sell your vision without giving away your blueprint."

He added that it's important to build an authentic community around your game and studio. "Because even if there's a copycat, they can't copy your audience, the community that is unique to you."

Metaxa agreed with Barnes' point, saying that it's important to build your studio's credibility so that the audience can expect a certain degree of "love, care, quality, and attention" from your games. But he added that the rapid explosion of AI clones has led to an abrupt change in industry thinking. "The traditional wisdom previously was to announce early, build up as many wishlists as possible over time, and then you'll have a good basis to launch your game eventually. And that seems to be changing fundamentally."

Ding said the advice Transcend Fund is giving to companies in its portfolio is now to think about the best time to release or make a big PR announcement, "and make that big push rather than thinking about doing it as early as possible."

Discoverability

The explosion of games made rapidly using AI is also deepening the problem of discoverability, particularly on Steam. Ding noted that the number of games released on Steam this October roughly doubled from the same time last year. "And I'm pretty sure that number will go even higher."

He added that even though the supply of games has "increased exponentially, people's attention hasn't necessarily increased." In other words, the number of games that find an audience has stayed roughly the same, as evidenced by the relatively static number of titles that receive 500+ Steam reviews. As Metaxa said: "It's much more intense competition for the pretty much same-sized pie."

The natural response to increased competition is to spend more time doing marketing, he said. "But then that gives competitors more time to just do fast copycats of your game if they're able to do it. And so it's kind of a Catch 22."

Metaxa noted that it "seems like Steam is moving in the AI direction", but that Roblox is already there, with the majority of games on the platform being created using AI. "Discoverability already on Roblox is nuts. Steam is getting more nuts. I have no idea how we're going to solve that."

Barnes added that the rapid shrinking of games media hasn't helped when it comes to getting word out. He estimated that the size of the games media has shrunk by between a quarter and a third over the past three to four years, "which is pretty devastating."

"I think that relationship that you have with press and creators is incredibly important too," added Metaxa. "They know your studio, they know the types of games that you make, and finding even those initial few creators that really believe in what you're making will make a huge difference."

"Discoverability already on Roblox is nuts. Steam is getting more nuts"

They explosion in the number of games must be hard for creators, too, he said. "They must be getting pitches absolutely non-stop."

Ding said that in such a congested environment, developers have to double down on building their audience. "From very early, day one, even before they start making the game, they need to start thinking about, okay, how can I start using a community? How can I talk to my players?"

The potential of AI

Although the rapid proliferation of games created using AI is causing headaches for developers and publishers, Ding said that the technology can have an upside.

"The true potential for AI is those AI-native games," he said. "It's about creating a new, innovative game experience that may not have even existed before today… That's the thing we are really, really excited about."

But Metaxa pointed towards issues with the current business model for games that rely on outside LLMs for their gameplay, singling out the Steam game Scam With Your Friends. "It went mega viral, [it was] super, super popular on Steam. But then the developer realised that they're having to pay for all of these tokens, and it's basically just eating into their revenue." Furthermore, Metaxa is sceptical of the use of AI in gameplay in the titles he's come across so far, saying that often it "seems kind of tacked on."

Barnes, however, highlighted how AI has been a boon for developers when it comes to exploring initial ideas or quickly creating early prototypes, particularly for "people who've got day jobs and they're not in the games industry yet, and they just simply wouldn't enter if it wasn't for the ability to make these prototypes and at least prove their own concept."

It could encourage them to pitch to publishers, with the intention to replace any generative AI assets later on, he said. "So getting that foot in the door, it can be really cool, because otherwise these guys would just simply not enter the industry."

"The true potential for AI is those AI-native games"

AI could also end up being used on the publisher and investor side. Ding said that in the last year, Transcend Fund has evaluated around 2,500 game opportunities, "which is a lot." The company is now trying to find ways to be more efficient when coping with the increasing number of pitches. "One thing we have been adopting is in-house AI tools that would help us to filter out something that definitely wouldn't fit. So that would give us more time to talk to the developers, rather than looking at things that are clearly off the mark." He gave the example of the tool automatically rejecting a pitch where the funding request was way above the threshold they would typically consider.

New platforms

Looking at game distribution platforms, Metaxa pointed to the big rise in web games in recent years, which are "bypassing the platforms altogether." He said that "pretty impressive games" can now be played through browsers. "I bet Steam are probably pretty worried about that, because maybe people won't need Steam in the future."

He singled out the web game platform Poki, saying it's "almost like doomscrolling on TikTok," but for games. "They're short, sharp, fun little games," he said, where the developers are paid according to how much time people spend playing their titles. "That seems to be a viable business model for smaller developers now, and is really starting to take off as an alternative for a particular type of game developer."

"I bet Steam are probably pretty worried about that, because maybe people won't need Steam in the future"

Ding also pointed to the rapid rise of mini games on the Chinese social network platform WeChat. "I can see potentially one day WhatsApp or other short message platforms might become that kind of equivalent," he said, adding that users will be able to discover games through messages from friends and then immediately play them on the same platform.

Ding also said that platforms that enable the creation of games using generative AI, such as Google Playground or Roblox, could open up development to more people. "Once you have a brilliant idea, it's now easier to implement that idea and make it happen," he said. "I can definitely see the potential."

Player reaction to AI

The news that House House used generative AI for a small portion of the development of Big Walk was setting the internet alight on the morning of the talk, illustrating just how divisive generative AI can be among players. Barnes said that Raw Fury makes a "big deal" out of publishing human-made titles, and that its audience is "not really interested in AI games."

But he does see the nuances: "I think it's important to separate AI into the creative tools and the process tools: I think there's a lot of acceptance towards the process tools," he said. "It's mainly the generative side that people have a big problem with, especially if it feels like it's replacing the human element: replacing an artist, replacing a writer, a voice actor, that kind of thing. I think in general, in our audience, they're really against it. I think for a more general audience, there's more of an acceptance. From a casual game perspective, [there's] kind of indifference, really."

"I think it's important to separate AI into the creative tools and the process tools"

Ding said it comes down to quality. He said that Transcend Fund has looked at player sentiment towards AI and found that not many players are against the use of AI outright, but are instead against the "low quality of whatever tools people are using." What they don't like is developers making less effort, he said, "while in the meantime players are still paying the same amount of money."

Metaxa agreed on the latter point. "If you're just doing a one line prompt and then being like, 'Oh yeah, I'm going to sell this', then no one really wants that," he said.

"The great thing about certain parts of AI is that you can tell that sometimes people put real love into crafting something as an individual that they may not have been able to otherwise on social media or something," he continued, "where they will take ages crafting some really specific vision of something that would have taken a whole movie studio to make, just for a meme. But you can tell that love and effort went into it. And for me, that's important: that someone is really putting effort into something, [and] has a vision that they want to show the world."

"I think if that kind of love goes into it, then it's meaningful."