I’ve seen Nvidia’s Project G-Assist AI assistant in person and I’m impressed: here’s why

Artificial intelligence is very popular in the tech industry today, and after seeing where it is going in games, I have more faith that it will one day replace strategy guides. It may still be a while, but I’m more convinced than ever that one day we’ll live in a world where you can ask a game for help and it will help you.

I saw my first glimpse of this potential future at Nvidia’s headquarters in Santa Clara; where I was recently invited to meet some of the team and check out some new tech demos.

G-Assist Project | Your AI Assistant for Games and Apps – YouTube

Watch

Easily the most engaging demo G-Assist Project, Potentially game-changing AI assistant from Nvidia Computex 2024 this summer. Even though it started somehow April Fool’s jokeProject G-Assist now looks like a surprisingly competent vision of a future where AI, not GameFAQ or Google, will answer your most pressing problems.

We won’t need strategy guides where we’re going

If you’re unfamiliar, G-Assist is Nvidia’s acronym for a set of tools developers can use to create context-aware AI assistants for their games. These tools can help developers take input from you as you play (via text, speech, or even your webcam) and feed it into AI models trained with various sources of game data so they can effectively answer your questions during gameplay.

The technology is still in development and not widely available, but Nvidia has partnered with Studio Wildcard to create a demo of how it might work in the company’s dino-filled survival game Ark: Survival Ascending.

(Image credit: Nvidia)

Nvidia has been using this demo to show off G-Assist over the summer, and now that I’ve tried it for myself, I can confirm that it’s impressive.

The vision of G-Assist that I see is basically a window that you can open in-game when you need to ask the game a question. You can then type or speak your question and (assuming you have a keyboard and/or microphone connected) the AI will respond with an answer within seconds using both Small and Large Language Models (SLM and LLM).

G-Assist can detect what’s on the screen and answer questions about what you’re looking at in-game, such as enemy weaknesses or what you can get out of them.

I also saw a live demo in which a video game character was shown an object in real life (via webcam), then successfully identified the object and answered questions about it. Game developers can control which sources of information models access. For example, the demo I tried used the Ark wiki, but the developers were able to limit resources to ensure certain information was hidden from players.

If this AI assistive technology lives up to Nvidia’s hype, you can already begin to imagine how strategy guides could start disappearing in 5-10 years. While I did notice some minor lags and occasional glitches during the live G-Assist demo, the technology worked as advertised most of the time.

(Image credit: Tom’s Guide)

There’s more than just AI game guidance here: Nvidia is trying to integrate this technology into games so characters can move and respond to your requests in real time. For example, I saw a demo of a game where an on-screen character could hear and respond to the player’s requests to switch between vehicles, optimize vehicles for different missions, and more.

And it can do more than just answer questions about how to succeed at your favorite video games. Nvidia also wants to talk about how this type of AI support can be accessed outside of a specific video game; so you can ask for help to dynamically optimize your computer’s performance for whatever you do that day.

For example, I saw a demo where an Nvidia representative asked Project G-Assist for help optimizing. Nvidia GeForce RTX 4090System equipped for efficiency. Right in front of me, I saw the AI assistant recommend locking the frame rate to 60 and undervolting the GPU to streamline performance and minimize fan noise.

As someone who gets tech support questions about laptop performance from readers and parents alike, this is huge. Performance enthusiasts will have plenty of room to argue about whether this kind of AI support is actually the best thing for your PC, but it’s easy to see a potential future that most people don’t care about: They can just click Tell the AI this. Optimize their PC for gaming and activate their favorite pastime.

Appearance

It’s no secret that Nvidia has been leaning heavily into AI for years, and now that I’ve gotten a glimpse of where it’s going in 2025, I’m intrigued.

Project G-Assist is still in its early stages, but the tech demos I’ve seen so far are intriguing and, frankly, a little scary. Currently G-Assist demos are a bit slow and clunky, but it’s easy to imagine these features coming to market in a few years. best computer games in the market.

If this future were to come true, it’s hard to imagine why you’d spend so much time searching for answers to your gaming problems on Google when you could just ask G-Assist. This might mean the end of the strategy guide business as we know it, but for people who want a more conversational relationship with their games, the benefits could be huge.