Post-Conference Blues and Headline Round-Up | 12/11/25
Sponsored by Lingotion: leading the way on legal and ethical AI acting for games.
The AI and Games Newsletter brings concise and informative discussion on artificial intelligence for video games each and every week. Plus summarising all of our content released across various channels, like our YouTube videos and in-person events like the AI and Games Conference.
You can subscribe to and support AI and Games, with weekly editions appearing in your inbox. If you'd like to work with us on your own games projects, please check out our consulting services. To sponsor, please visit the dedicated sponsorship page.
Hello one and all and welcome to this week’s edition of the AI and Games newsletter. Apologies for missing you inbox last week, but after being in London from Saturday to Wednesday prepping, running, and tearing down our very own AI and Games Conference 2025, I was running on fumes. I had naively thought I could get an issue out and alas I fell prey to my own ambition - or more rather I just went to bed and slept, like a lot!
This week I figured I’d share some of my own personal thoughts on how things went down with the conference, but also we need to catch up with the news. A whole bunch of stuff has gone down in recent weeks - most of it not great tbh - and it warrants a bit of Scottish sensibility being thrown at it.
As the wee Italian man likes to say, lets-a go!
Follow AI and Games on: BlueSky | YouTube | LinkedIn | TikTok
A Word From Our Sponsor
We’re grateful once again to the good folks over at Lingotion, who are not just leading the way for legal and ethical AI acting in games, but sponsoring this weeks issue!
We’ll have a sit down conversation with founder Andreas Rodman later this month to discuss how Lingotion provides a legal and ethical framework for AI voice acting that gives performers legal control and renumeration. In the meantime, be sure to check out Lingotion’s platform and their approach to supporting actors and performers through AI at lingotion.com
AI and Games Conference 2025!
I find it hard to believe it’s been a week since the 2025 AI and Games Conference. I’m still recovering and largely very, very, tired. But I am absolutely thrilled with how things went down.
As I expressed over on LinkedIn last week, the goal for this year’s event was to prove quite simply that we could do it again. We wanted to build on last years conference to create a larger event, and cultivate a space where expertise takes precedence over the hype that is flooding our sector, where our values place an emphasis on practical developments and the people that make it happen, and that we continue to build an accessible and relevant space for people who work the game AI field.
Thus far the sentiment I’ve received from attendees, speakers, and sponsors, is that we have achieved that and then some. Over 30 presentations across two days from the leading experts in the sector, bringing together game AI programmers, academic researchers, leading innovators, and some additional outside expertise in investment and legal. There really is no other event like this at the moment - it’s why we built it to be just like this!
There is much to be done in the coming months, as we not only put together the video recordings of the programme (which will go online in 2026), but our post-event survey will go to attendees shortly as we collate your feedback. Plus a video summary will go live on main AI and Games YouTube channel in December. We know there is still a lot of room for improvement, but I think we’re getting the key things right: great presentations from great speakers, great vibes, and a space to bring the community together. To that end, we are already working on events planned for 2026 and 2027, and will have more to say on that in the new year.
In the the meantime, a huge thank you once again to everyone who has supported us. Be it as a sponsor, as a speaker, as an attendee, our volunteers in those very green t-shirts, or has shared their appreciation for our work online and anywhere else. It’s the community we’re building that makes this event work, and it’s important to my co-founders and I that we continue to deliver not just an event that reflects game AI at its best, but delivers on what our community needs: the opportunity to exchange ideas, to network, and build connections that are practical, relevant, and valuable.
Thank you once again to everyone involved in running the conference, who continue to be just a fantastic group of people to work with. What makes this that much easier is we have such a talented, excited, intelligent, and empathetic group of individuals making it happen. We’re just getting started, and I’m excited to share what is coming your way in the future.
AI (and Games) in the News
Alrighty, here’s some headlines that caught my eye in the past couple of weeks.
Meta Cuts Jobs in their Superintelligence (AI) Labs
As reported over on Reuters in late October, Meta (y’know, Facebook, Oculus, WhatsApp, Instagram) have laid off 600 people from the thousands hired across their superintelligence labs - essentially the umbrella term given to all of their teams and orgs that are focussed on building next generation AI systems such as their Llama large language model (LLM). Like many others, they’re struggling to turn this into a significantly profitable venture, with the likes of the Llama 4 model receiving a rather poor reception on launch back in April.
Arc Raiders Storms Out the Gate, but Generative AI Adoption Sours Reception
Embark Studios new extraction shooter Arc Raiders has had a very successful launch, setting a concurrent player record that exceeds my beloved Helldivers 2, and it doesn’t hurt that most reviews have identified that it is really fun to play.
I say most reviews given it’s 85%+ Open Critic rating was soured by Rick Lane’s review on Eurogamer that - to my knowledge - is the only one of the larger gaming sites that called out Embark’s ongoing use of Text-to-Speech (TTS) generative AI technology for the in-game voice work. This has since led to the discourse blowing up on social media in the past 48 hours.
Now it’s worth noting that Embark doesn’t really clarify that they used generative AI, with the Steam generative AI disclosure providing this nothingburger of a statement:
During the development process, we may use procedural- and AI-based tools to assist with content creation. In all such cases, the final product reflects the creativity and expression of our own development team.
I find it rather duplicitous of them to obfuscate how extensive the generative AI usage is on the store page. Particular when you note they were more explicit in their acknowledge of TTS technology for the announcer voices in their previous title The Finals. Though it has been discussed elsewhere that these are voice models trained on human actors.
I haven’t had a chance to catch up with this myself - give y’know, I just ran a conference - but I’m keen to look into this, given I suspect I will have an opinion or two on this to share.
Square Enix Wants Generative AI to Handle 70% of QA by 2027
Every now and then I read a statement where I think ‘okay, uh-huh’, and this is certainly one of those. As reported on Eurogamer, Square Enix announced this idea in their recent progress report to investors as follows:
Joint Development of Game QA Automation Technology Using Generative AI with the MatsuoIwasawa Laboratory, the University of Tokyo
[Goal]
Automate 70% of QA and debugging tasks in game development by the end of 2027. Through the use of automation technology, aim to improve the efficiency of QA operations and establish a competitive advantage in game development.[Structure]
A joint research team of more than ten members, comprising researchers from the Matsuo-Iwasawa Laboratory UTokyo, and engineers from the SQUARE ENIX group, is advancing the collaborative development.
It’s also worth noting that their intent to explore ‘overseas structural reform’ mentioned in the same presentation is what lead to recent job losses in Europe and North America.
Now as Eurogamer stated in their headline, Square Enix are notorious for chasing trends, with their ill fated attempts at Metaverse’s and NFTs going down like lead balloons, and so it seems only natural that generative AI was the next big thing for them to tackle. In fact I pointed this out last year when they said in their new year letter for 2024 that they would be “aggressive” in their pursuit of AI.
But AI at Square Enix is nothing new. They’ve been doing this for a long time, and utilise both traditional game AI for NPCs, but also have really embraced machine learning for aspects such as gameplay balancing. Square Enix joins many studios who already use some form of AI for QA purposes, notably Ubisoft and EA, and while I don’t have any clear insight I imagine they’ve been doing this already for years.
But I think a lot of this stems from two key points:
Yes they already use AI for QA but what impact this has on workers is a big issue. In some cases, having AI do 70% of the QA is actually useful, given it lets QA teams focus on aspects that their (human) skills are better suited for. After all there’s a baked assumption that they’re saying 70% of the current amount of ‘effort’ in QA, whereas it’s possible human QA hours don’t change but you start to achieve more coverage through AI. So if this is implemented in a way that makes sense, and protects workers, that is… good, right? Of course, take that perspective with a massive pinch of salt.
In the same presentation they mention their intent to “Promoting AI Utilization in Japan”. This is really important. Given you can read this whole news story as an intent to show to Japanese businesses, customers, and governance, that they are invested in the space. This ties into my points from a few weeks back about the “AI Promotion Act” published last summer in Japan that encourages businesses to show they are embracing AI in ways that align with the broader national agenda. This was important given Nintendo uncharacteristically spoke up to deny allegations of lobbying the Japanese government on AI adoption. Given it looked like they were going against the grain of the government.
So in truth I read this as mostly an effort to placate Japanese government more than anything else. Even if the rest of the games industry, and consumers, roll their eyes in the process. Now, sticking with the Japanese AI discourse…
Japan Trade Representatives Demand OpenAI Remove Respect Their IP
As I discussed back in March, whenever OpenAI releases new models, you’ll more often than not see people use them to exploit Japanese intellectual property such as Manga and Anime. When GPT-4o released earlier this year we had memes restyled to reflect Studio Ghibli animation, and the same happened again not that long ago as Sora2 came out as well. This is because, referring back to my issue on Nintendo denying lobbying the government, there aren’t laws that protect IP infringement from AI in Japan. Rather, they operate on more of an honour system, whereby companies are named and shamed publicly for their misdeeds. A concept that largely works in Japan given their corporate culture, but doesn’t really translate to the wider world. Plus it doesn’t help that OpenAI clearly have no morals whatsoever, so hey… what’s to be done about it?
The Content Overseas Distribution Association (CODA), which operates on behalf of the Japanese Ministry of Economy, Trade and the Industry, and the Agency for Cultural Affairs, issued a statement declaring they have written to OpenAI at the end of last month regards Sora2’s exploitation of Japanese IP such as games, anime and manga.
As stated, they have requested that CODA’s members content not be used without permission, and for OpenAI to respond to inquiries regards the usage of copyrighted materials.
I suspect OpenAI will do nothing of substance to respond to this, but it’s interesting that the Japanese government is beginning to be frustrated by the exploitation of one of their largest and most lucrative exports. After all, CODA’s membership include Aniplex (Demon Slayer), Bandai Namco (Tekken), Studio Ghibli (Spirited Away), Kadokawa (one of the largest publishers of manga) and Shueisha (Shōnen Jump).
“You End Up With Something Pretty Derivative”
Y’know every now and then Take Two CEO Strauss Zelnick says something where I think actually, he’s alright. In a recent interview at CNBC’s Technology Executive Council Summit, he responded to a query of whether generative AI could make a future entry of GTA, and gave less than a positive response.
Critically this sentence stuck out:
Whether people in Silicon Valley like to hear this or not, AI is big datasets, with lots of compute, attached to a large language model. So what are datasets by definition? They’re backward-looking.
On one hand his view of AI is a tad reductive, but if you listen to the full response I largely agree with him.
Of course days later story broke about Rockstar staff layoffs in what appears to be an effort in union busting, so yeah… sorry pal we’re back to square one.
EA Partners With Stability AI to ‘“Reimagine Game Development”
I don’t have much to add here, except:
We talked about this already, so it didn’t surprise me at all.
As PR statements go, it’s impressive that you can write over 500 words and somehow say fuck all by the end of it.
By comparison, this issue is roughly 3000 words of jam-packed goodness!
This push into generative AI is, surprisingly absolutely nobody, not going down well inside EA.
Krafton Declares Intent to Become an “AI-First Company"
Okay we’re just making stuff up now. Long story short, they’re all in on generative AI. I had heard some rumblings of this from a meeting I had with Krafton at Gamescom, but outside of general discussions of just embracing AI - and investing a ton of money into GPU clusters - it doesn’t give much insight.
Bringing Law & Order - With Generative AI
Yeah, this story cropped up over at VGC. Wolf Games - founded by son of Law & Order producer Dick Wolf - has signed a deal with NBCUniversal to make games based on NBCs IP (which includes Law & Order, but also Chicago PD, Sex and the City, and SNL).
You can see perhaps what they envisage from the teaser Wolf Games put out back in March for their AI-generated game ‘Public Eye’. Personally I think it looks really rough. Notice how none of the generated video clips last longer than a couple of seconds - because at this time they can’t without it losing cohesion.
I mean at least they still spend money on actual real human episodes of Law & Order, largely to pay Ice-T’s retirement fund.
Say Hello to My Litt - Oh No, Wait…
This one I just had to bring up because it’s just baffling.
As reported on various sites, but I’ll point to VGC again, an store page cropped up last month for the 2006 videogame adaptation of the 1984 movie Scarface: The World is Yours on Steam and the Epic Games Store. This was largely unexpected, and it was being hosted by a company called EC Digital. The new store page suggested a version built with upscaled graphics (achieved using AI) and that the game would be rebuilt to run on modern hardware. However, they accidentally published the game, and for a brief time it was available to buy on Steam.
Savvy consumers then discovered it was just the original version of the game (it was pulled from stores years ago), but simply full of mods that people have built for the game over the years - and included without their permission. Plus it became apparent that the rights to the game had not been signed off, and long story short EC Digital had essentially went and smacked this bundle together as proof they could publish the game before they pitch to the game and movie rights holders.
The mysterious publisher have since published the statement above and then gone dark. It seems likely that AI-enhanced remaster is not going to happen.
Wrapping Up
Thanks for sticking through and getting to the end of this weeks issue. Things will be returning to normal in the next couple of weeks, and we kick things off with an interview with the folks at Meaning Machine - the small UK-based indie outfit building their own conversational AI pipeline for AI-native games.
I sat down with Ben Ackland and Thomas Keane a couple weeks back before they took the stage at our AI and Games Conference, and discussed the lessons learned over the past couple of years. It’s going to be an interesting one, join us next week for the deep dive!