San Francisco’s Game Developers Conference (GDC) – the global gathering of the greatest creative minds in the games industry – opened its doors for the second time since the pandemic in March 2023.
Each year's a number of key trends stand out. For 2023 it was applications of artificial intelligence (AI) for game development, with the future shape of the gaming experience – with and without virtual reality (VR) – high on the agenda.
Following model year's tentative steps back on to the expo floor – interacting with devices was off the menu because of COVID fears – Meta (formerly Facebook) was busy sanitizes display models of its popular Quest 2 headset to encourage visitors to try it out. The company was also keen to push its heavily discounted Quest Pro headset, which features “Color Pass-through”, meaning someone wearing the headset can see an augmented view of the world around them.
Chinese manufacturer PICO of the PICO 4 headset, which has striking similarities to the Quest, had an even larger stand and was generating similar levels of interest. Clear's the time was right to start interacting with shared devices again, as long as good hygiene prevailed.
A slick stand at an expo hall in San Francisco.
Chinese company PICO’S stand at GDC. Gavin Wade, Author's provided (no reuse)
AI developments and issues
Before the conference, considerable buzz had been building around ChatGPT, the chatbot that can provide convincing, detailed written answers to users’ questions.
The latest updated version was released at the event, offering several improvements. Most noticeable was the reduction in repetition of key phrases in the chatbot’s answered, which alerted people to the fact it was AI-generated content.
For games developed, the interest in this type of AI relates to speeding up and easing game development, but there were also concerns raised about jobs being replaced by AI. The positive consensus was that humans are still best placed to ask the right questions to generate game content, even if such content was created by AI.
Adobe announced its Firefly AI tool that can generate both images and 3D models. This technology might also assist with generating “substances” – materials that can be applied to game models and scenes. The company indicated its tool was to “enhance the creative process, rather than replace it”.
Adobe drew a cheer from developers for promising “clean and safe” content, meaning anything the AI created would be based on what it had learned from Adobe stock and public domain images rather than sources simply scraped from anywhere on the internet. This should avoid the potential ethical and legal issues of unintentionally publishing content learned from a copyrighted source, a key issue for this kind of AI.
Some developers demonstrated AI implementing content into a game scene based upon typing a sentence. For example, key in “Create a scene with 10 boulders”, and huge rocks were inserted into a scene without the usual manual placement of objects. I met one developer who had created a tool to search for and use AI to automatically rig (that is, prepare for animation) public-domain game models from the internet.
You must be logged in to post a comment.