2023 Technology Snapshot
Important Note: This article reflects the state of AI technology and tools available in December 2023. The AI landscape has evolved significantly since then. Throughout this article, you'll find [2026 Update] boxes highlighting major changes and current alternatives.
This retrospective captures a pivotal moment in AI-assisted game development and serves as a historical reference for how quickly the field progresses.
Introduction
Today, artificial intelligence plays a pioneering role among rapidly evolving technological trends. This dynamic and innovative technology is also increasing its impact in the gaming sector, manifesting itself in various fields both directly and indirectly.
This article focuses on the wide range of applications of artificial intelligence in the game industry, emphasizing how game developers and industry professionals can benefit from this technology.
We'll also explore various AI-powered platforms that game developers could use in 2023: ChatGPT, Bing, Bard, Inworld AI, and DALL-E 3. These platforms assisted developers with code suggestions, image generation, character creation, and much more.
Enemy Behaviors and NPC Intelligence
Programming enemy behaviors plays a crucial role in the game development process and is achieved using artificial intelligence techniques. AI can be used to make enemy characters behave more realistically and intelligently, enabling them to respond better to player strategies and actions.
We call these NPCs (Non-Player Characters). Here are some AI techniques and examples used to make enemy characters behave more intelligently:
Core AI Techniques for NPCs:
- Pathfinding - Pathfinding algorithms enable enemies to find the shortest or most suitable route
- Speed and Acceleration Control - Adjusting enemy speeds and acceleration to make movements more natural
- Salvo Attacks - Enemies can attack in salvos according to specific strategies, then calm down
- Group Attacks - Enemies can cooperate as a group to attack the player
- Field of View - Enemies have a specific viewing area to detect players
- Sound Perception - Adding sensitivity to sounds made by players to determine enemy location
- Escape Strategies - Enemies can use escape or hiding strategies when below certain health levels
- Cover Usage - Enemies can use their surroundings to avoid attacks or defend themselves
The Common Goal
All these techniques share one common goal: to mimic real player behavior. The more convincingly NPCs act like human players, the more engaging and challenging the game becomes.
DLSS and FSR: Machine Learning for Graphics
Compared to other technologies, DLSS (Deep Learning Super Sampling) and FSR (FidelityFX Super Resolution) attract more player attention and both leverage artificial intelligence.
These technologies fundamentally use advanced machine learning to significantly reduce the load on graphics cards. They generate additional frames between existing frames through a series of predictions, filling the gaps.
Technology Comparison:
- DLSS - Released by NVIDIA, utilizes deep learning neural networks trained on supercomputers
- FSR - Released by AMD, uses spatial upscaling algorithms without requiring ML hardware
Especially with RTX 4000 series graphics cards, DLSS 3.0 provides incredible performance gains, sometimes doubling or tripling frame rates while maintaining visual quality.
2026 Update: Graphics Upscaling Evolution
DLSS has evolved to version 3.5+ with even better image quality and frame generation. Intel has entered the market with XeSS (Xe Super Sampling), providing another hardware-agnostic alternative. FSR 3.0+ now includes frame generation comparable to DLSS. These technologies have become standard in modern AAA games, with most engines offering built-in support.
Narrative and Storytelling
Artificial intelligence plays an important role in game development processes, enabling the automatic creation of game worlds, stories, and character interactions.
Procedural Content Generation (PCG):
Using PCG, large and dynamic game worlds can be created. AI supports dynamic storytelling by adapting to player decisions and events. Interactions between characters can be made more realistic with decision trees and dialogue systems.
Personalization and Emotion:
Game dynamics can be analyzed by AI to make the player experience more personal. Additionally, sentiment analysis can understand player reactions and increase emotional richness in storytelling.
AI can improve the game's replayability by adding random events and quests, allowing the game world to be richer and more diverse. These features provide game developers with the opportunity to offer creative and interactive gaming experiences.
AI Tools in 2023
In 2023, several AI platforms emerged as essential tools for game developers. Let's explore the major players and their capabilities at that time:
1. ChatGPT
No need to even mention it—when OpenAI released ChatGPT, it stuck in everyone's vocabulary, from A to Z, even those who didn't fully understand what AI was. Within a week, hundreds of pieces of content were produced on platforms like YouTube about it.
While serving many purposes, ChatGPT was particularly useful for game developers in providing code suggestions and determining game stories. Additionally, ChatGPT 3.5 provided advice during and after your game development and notified you if there were errors.
Important Caveat (2023)
It's worth noting that it doesn't always provide 100% accurate information. Also, if you intended to use ChatGPT 3.5 for free, unfortunately it was only current up to January 2022. However, in my personal opinion, this didn't pose a major problem.
2026 Update: ChatGPT Evolution
ChatGPT has evolved to GPT-4 Turbo and GPT-4o with significantly improved accuracy, real-time information access, multimodal capabilities (text, images, audio), and much larger context windows. Specialized models like ChatGPT Code Interpreter and GPT-4 Vision are now standard. Many game studios have integrated GPT APIs directly into their development pipelines.
2. Bing AI
Bing, a Microsoft product, uses ChatGPT 4.0 infrastructure. It's free and helps with many tasks. Because it's backed by ChatGPT 4.0, it naturally does many things that ChatGPT 4.0 can do. Additionally, since it also uses Bard and DALL-E 3, it can generate images for you.
2026 Update: Microsoft Copilot
Bing AI has been rebranded as Microsoft Copilot and is now deeply integrated into Windows 11, Microsoft 365, and various development tools. It uses GPT-4 and DALL-E 3 with enhanced capabilities. Copilot is now a major player in coding assistance and game development workflows.
3. Bard
Google's AI was at least as powerful as ChatGPT in 2023. However, possibly because it was still experimental, unlike Bing's use of DALL-E 3, Bard didn't draw images but instead found images from the internet that matched your criteria. So I don't know how suitable Bard was for your needs in this regard.
2026 Update: Bard → Gemini
Major Change: Bard no longer exists! It has been replaced by Google Gemini, which is significantly more powerful. Gemini now offers:
- Multimodal understanding (text, images, video, audio, code)
- Native image generation capabilities
- Better coding assistance with Gemini Code Assist
- Integration with Google Workspace and Android Studio
- Gemini Ultra, Pro, and Nano models for different use cases
4. Inworld AI
Inworld AI was a platform used to create AI-powered characters. It allowed developers to create, customize, and manage their own virtual characters.
Inworld AI used large language models (LLMs) to create character personalities, thoughts, memories, and behaviors. LLMs were trained on large datasets of text and code, ensuring characters were realistic and engaging.
Inworld AI used game engines like Unity or Unreal Engine to create character animations and interactions, ensuring characters offered an interactive and immersive experience.
2026 Update: Advanced NPC Systems
Inworld AI has matured significantly and now powers AAA game NPCs with real-time conversation, emotional intelligence, and persistent memories. Competitors like Convai and Character.AI have also emerged. Many game engines now offer native LLM-powered NPC systems. NVIDIA's ACE (Avatar Cloud Engine) provides similar capabilities with on-device inference.
5. DALL-E 3
DALL-E 3, released by OpenAI (creator of ChatGPT), was sometimes more successful than other image-generating AIs on the market. Of course, there were very powerful alternatives to DALL-E 3, including: Midjourney, Stable Diffusion, Starryai, and NightCafe.
2026 Update: Image Generation Revolution
The image generation landscape has transformed dramatically:
- Midjourney V6+ - Industry-leading quality with photorealistic capabilities
- Stable Diffusion 3+ - Open-source with incredible customization
- Adobe Firefly - Commercial-safe AI integrated into Creative Cloud
- DALL-E 3 - Still competitive, integrated into ChatGPT Plus
- Runway, Pika Labs - Now generating videos, not just images
Real-time AI art generation is now possible on consumer GPUs. Many game engines offer native AI asset generation tools.
Conclusion: The AI Revolution in Game Development
While AI has become important in every field, it holds even greater significance for us game developers. We developers must take full advantage of this ever-evolving AI technology.
In addition to popular platforms like ChatGPT, Bard, and Bing, we should leverage other successful AI tools to push the boundaries of creativity and make future gaming experiences even more exciting.
Reflections from 2026
Reading this article from a 2026 perspective highlights how rapidly AI has evolved in just three years. What seemed cutting-edge in 2023—ChatGPT 3.5 with 2022 data—now feels almost quaint.
Key changes since 2023:
- Multimodal AI (text + images + audio + video) is now standard
- Real-time AI in game engines is commonplace
- AI-generated game assets have become production-ready
- LLM-powered NPCs offer truly dynamic conversations
- Procedural generation has reached new levels of sophistication
The principles discussed in this article remain relevant, but the tools and capabilities have advanced exponentially. This retrospective serves as a reminder of how quickly technology evolves in the AI age.