Level Up: AI in Gaming

Generative AI is streamlining game development and enhancing in-game experiences for players.

Game building is typically resource-intensive in nature. Gen AI can significantly reduce this by generating 3D assets, virtual environments, in-game conversations, game testing bots and more. Startups like Luma AI, Common Sense Machines are enabling the creation of 3D assets from text / image prompts. Gen AI tools like Inworld AI are already gaining traction amongst developers and user-generated content (UGC) creators – Inworld AI has been used to create mods for popular games like Skyrim and Grand Theft Auto V.

Large game developers are taking notice – Microsoft's announced a collaboration with Inworld AI to design a multi-platform AI toolset, including an AI design copilot for scripts and dialogue, and an AI character engine that can be integrated into games and used to dynamically generate stories, quests, and dialogue.

However, major publishers have differing opinions on AI-generated gaming content. While Steam removed certain games containing AI-generated content (citing copyright risks), Epic Games has welcomed games containing AI-generated content.

The gaming toolscape is buzzing with activity - innovation across different stages of game development, strong consumer interest, and fundraising momentum. Read on for an overview of the market.

Use Cases: Gen AI significantly reducing time and cost of game production

Cost-effective, time-efficient production: The game development process is time-intensive and costly, and often involves expensive stock assets from stores like Unity’s. Gen AI can create assets with just a prompt, cutting down on both time and resources required. Luma AI allows developers to create basic 3D characters to complete interactive 3D scenes by entering text prompts, even on their mobile phone. Scenario is helping game studios like LLC create thousands of assets monthly for 10% of the cost, compared to their earlier capacity of only a few assets per month. Gen AI can enable individual developers and smaller studios to be creative as well as cost- and time-efficient in production.

Better NPC conversations, responsive game environments: Gen AI can modify game environments on the fly to adapt to player choices and generate NPC barks (dialogues from non-player characters) based on gamer activity. Inworld’s AI Engine can respond to players' actions and decisions with dynamic game states, environments, and AI agents – transforming standard gameplay into a bespoke experience for each player, each time they play.

User-generated content (UGC) creation becoming more accessible through Gen AI: User-generated content (UGC) and modding (players modifying aspects of a game) have been integral to the gaming fandom, and led to popular games like Dota 2 (Warcraft III) and Counterstrike (Half-Life). Players are able to create in-game assets (like custom maps, items, skins) or entirely modified versions of the game that they can share, or sell to other players – making games larger and more personalized. Companies have often incentivised players to create UGC – Epic announced that they’d give 40% of Fortnite’s net revenue back to its UGC creators through engagement payouts.

Gen AI enables anyone who’s interested to create game content & assets without needing special design / coding skills, thus reducing the skill barrier. This will make UGC creation more accessible and abundant. Roblox, one of the largest platforms for UGC creators, has announced an AI Assistant which lets creators build virtual assets and write code with Gen AI.

Multiple game developer use-cases for Gen AI

• Storyline & content: Ludo AI generates a variety of game concepts on the basis of keyword prompts, and further develops and refines the selected concept.
• 3D Asset Generation: Luma AI enables the creation of 3D assets and interactive scenes using an iPhone or through the web.
• Sound Effects: ElevenLabs generates sound effects for video games from a text description.
• Motion graphics: Move AI brings realistic human motion to animated characters by using 2D videos to capture motion data for 3D characters.
• Game testing: Agentic deploys AI agents as live players of video games to identify issues in gameplay.

Alongside assisting developers, AI is also being used as a gamemaster in character-based games like AI Dungeon, The Fortuna, and Gandalf by Lakera, fully directing the game's progression and interactions.

Market Map: Active AI innovation around dev tools; asset creation and AI assistants are focal areas

AI innovation is largely centred around game developer use-cases, especially asset creation

AI innovation in gaming is heavily focused on developer tooling, especially around tools that automate the creation of game assets. Asset creation is traditionally tedious and labor-intensive, and purchasing assets from stores is expensive. Tools like Luma AI, Hypothetic can generate detailed 3D characters, objects, textures, and entire visual worlds from simple text prompts. AI asset creation tools solve for one of the biggest struggles in game dev – having to compromise between creativity and efficiency.

Like other Gen AI applications, the Copilot model is being explored in Gaming as well by large game dev platforms

The Copilot model, already popular in coding and copywriting, is gaining traction in gaming. Industry incumbents like Unity and Roblox are coming up with AI Assistants that support multiple stages of game development. AI copilots that are capable of creating complete games from text prompts are coming up — crafting everything from immersive worlds to characters and narratives. Some startups offering AI copilots for game development are Rosebud AI, Series AI, and Astrocade.

Funding Landscape: Active funding in 2022-24 despite slowdown, including strategic investments

Robust funding landscape with 30+ deals for emerging startups in 2022-24

Despite a 72% decline in overall gaming investment in 2023, AI gaming startups have attracted several deals, primarily Seed and Series A rounds. Both VCs & strategic investors are actively participating in the market:
 

• BitkraftVC, founded by Jens Hilgers (former CEO of ESL) is heavily invested with 7 deals, including companies like Inworld AI, Latent Technology, Volley, and Series AI.
• Prominent strategic investments include NVIDIA’s investment in Luma AI, Disney’s investment in Inworld AI, and NBCUniversal's investments in Inworld AI and Charisma AI.

Inworld AI is the highest funded startup in AI x Gaming

Inworld AI’s $50M deal in 2023 made it the highest funded startup in the space. This also put Inworld AI at "halficorn" status, with a post-money valuation of $500 million.

Tool Popularity: Inworld AI, ConvAI have received strong developer traction and feedback

The game developer community builds on game engines like Unity, Unreal Engine, thus, several dev tools are listed on the asset stores of these engines. Reviews on these stores serve as a good measure of a tool’s popularity.

Both the top popular choices are serving towards better in-game NPC conversations.
 

• Inworld AI has been used by developers to create AI-driven character mods for Skyrim, Stardew Valley and Grand Theft Auto V.
• ConvAI uses Gen AI to enable non-player characters (NPCs) to engage in realistic conversations and independent decision-making. ConvAI powers the conversations on 3D virtual world / metaverse Second Life.

Web Traction: Luma AI seeing exceptional traction; CSM & Inworld AI gaining interest

Luma AI uses AI to enable users to generate realistic 3D environments from text prompts and 2D photographs. The process is entirely doable on an iPhone via the Luma AI App. This tech has attracted interest from a broad audience including game developers, architects, VFX professionals and video content creators. Luma AI is backed by NVIDIA.

Common Sense Machines (CSM.ai) offers end-to-end 3D worlds generation from generating 3D assets, to animation, to stylization. Users can generate 3D models from text prompts, sketches or 2D images.

Inworld’s Character Engine uses AI to generate NPCs whose personalities, thoughts, memories, and behaviours are designed to mimic the social nature of human interaction. Inworld has worked with multiple AAA brands including Disney, NetEase Games and Niantic to build AI-driven NPC experiences. Inworld is also Microsoft’s official partner for a multiyear project to enhance Xbox development tools, including an "AI design copilot" for developers to create AI-driven characters, narratives, and quests.

Looking Forward

1. Robust AI innovation in game dev tools, will reduce tradeoff between creativity & efficiency
   Gen AI is revolutionizing game development by providing tools that automate several aspects from asset creation (Luma AI, Hypothetic) to production (Inworld AI, Rosebud AI) to gameplay (Latitude). Gen AI will enable game developers to expand creatively without being constrained by resource limitations.
    
2. Player experience in games will significantly change for the better
   The gameplayer will be one of the biggest winners in AI x gaming. Gen AI tools like Inworld AI, ConvAI are already being used to give players better social interactions and responses from NPCs. Future games are likely to have more immersive, personalized experiences for players that adapt to their in-game actions and choices, and change every time they play – making the game world feel more alive and socially rich.
    
3. User-generated content (UGC) and mods will become a thriving economy around popular games
   Gen AI greatly reduces the skill barrier required for game design and development – allowing any gamer to easily create UGC like in-game assets, maps, costumes as well as modified versions of the games. An increase in UGC will lead to larger, more personalised games – game engines / publishers like Roblox, Epic recognize this and are actively incentivising players with toolkits and monetary benefits to create content.
    
4. AI gaming is attracting investment, expect breakout tools to emerge soon
   Emerging startups are seeing active interest from VCs and strategic investors. While the current investment landscape has 30+ speculative bets; the real standout tools will become apparent as wider adoption highlights which are most likely to succeed, leading to bigger funding rounds. Expect Inworld AI, Luma AI and ConvAI to raise future rounds. Inworld AI is a likely candidate for unicorn status soon.
    
5. Fair-use and copyright issues yet to play out, similar to AI video / images
   Similar to AI content creation across image, copywriting and video, the use of Gen AI in gaming comes with its own share of fair-use debates. Steam has removed games for containing AI-generated content, citing copyright risks. Videogame voice artists have received SAG-AFTRA approval to strike if a favourable contract doesn’t come through. This conversation will intensify and shape how major studios and publishers act in the wake of an increasingly AI-native gaming landscape.