The Final Wave of Speakers at the AI and Games Conference 2024
More Horizon Forbidden West, Automatic Playtesting, AI and IP, and the Warhammer 40K: Space Marine 2 AI Post-mortem
With just over two weeks to go until the inaugural AI and Games Conference, it’s time to announce our final batch of talks. You can catch the posts we released over the last two weeks highlighting everything from the AI of Rainworld, Aria: History Untold, the use of LLMs for issue tracking in Battlefield, Unity’s Muse platform, Unreal’s Learning Agents and more.
Today we’re rounding things out with some absolutely killer content! Including more about the AI of Horizon Forbidden West, automatic playtesting of board games, the risks to your intellectual property when using generative AI, and the AI post-mortem of one of the most talked about titles of the year: Warhammer 40K: Space Marine 2.
Tickets and Recordings
As a reminder: our talks will be recorded for everyone else to catch online in early 2025. Meanwhile the final batch of tickets went live at the end of last week and very quickly sold out. We appreciate your patience and no doubt some will be disappointed in missing out. The response and interest in our event has vastly exceeded our expectations for this inaugural event, and we’re certainly taking this into consideration for if we run a follow-up in 2025.
Our Final Speakers
It’s time to welcome our final batch of speakers to this years conference! But before we do, one final element of the event we will be our ‘Future of AI for Games’ panel that we will be hosting during the afternoon on November 8th. We can confirm that Yassine Tahi (CEO of Kinetix) and Roberto Lopez Mendez (Director, Tech Lead Developer Advocacy at Arm) will be on stage, with more panellists to announce next week!
Now, let’s get onto the final set of speakers: delivering even more fantastic talks to add to the line-up!
HTN Planning in the Decima Engine
Guerrilla Games, the creators of the Killzone and Horizon franchises, have long invested in the use of Hierarchical Task Network (HTN) planning for the behaviour of their non-player characters. In this talk Tim Verweij gives us a deep dive into the underlying architecture they’ve built for HTN in their proprietary game engine ‘Decima’.
Empowering Game Designers with Automatic Playtesting
Building a new board game can be quite risky, especially when you need to playtest it significantly in order to figure out how to move it forward. Raluca Gaina and Diego Liebana are AI researchers based at Queen Mary University of London, and are the founders of Tabletop R&D - a new venture to facilitate automatic playtesting of board games. In this talk they provide us an overview of the AI processes they use, and the benefits this provides to game designers.
Escaping the Infinite Mid
There has been a lot of talk in the past couple of years about how Large Language Models (LLMs) will change the types of experiences players will have when working in games. But thus far, we’ve seen little that could be ready for public consumption, there are lots of issues in tailoring these systems to get the best value from them. One of the leading companies in building LLMs for games is Hidden Door, and we welcome Chris Wallace - one the companies engineers - to share his experiences and guidance on how best to navigate this challenging space.
Navigating AI and IP: A Practical Toolkit for Devs
One of the many issues that developers will face when using LLMs and other generative AI tools is protection of your intellectual property. To help give a clear and concise overview of how to navigate these issues, we welcome Anna Poulter-Jones, a games lawyer for legal firm Sheridans to clear the air, and set the record straight!
RL Agent Training is Property Based Testing
Jaden Travnik is an engineer at Sony AI whose experience in training agents for games using reinforcement learning is being brought to the fore in this talk. The truth is that training RL agents is often challenging during development, with unintended functionality arising. As such, Jaden will talk about how we can use RL to more effectively catch issues as they arise as part of property based testing.
Human-level AI Bots in Soccer Kids by Acid Wizard
Soccer Kids is a unique tactical sports game where players have to plan for future actions, while also handling the movement of physical objects - notably the football! In this talk Rafal Tyl from QED Games goes into detail on how they adopted Monte-Carlo Tree Search (MCTS) to address these challenges!
Space Marine 2: AI Post-Mortem
With a stack of high-quality releases over the past couple of years, we’re most certainly in a new golden era of Warhammer 40K video games! Space Marine 2 - the sequel to the cult classic from 2011 by Relic Entertainment - sees the return of Captain Demetrian Titus of the Ultramarines to wage war against the Tyranids. We welcome Vladislav Iantsevich from developers Saber Interactive to give us an insight into the challenges faced in building the game, and how they went about resolving them.
Just Over a Week to Go!
Oh boy, time has flown by since we announced this event but three months ago, and we can’t wait to see some of you on the day in London. For everyone else, we’ll keep you posted in the coming weeks as to where and how you can catch these talks online!