Meet your new and improved teammate! Human-directed AI agents and gaming
summary
Claim: language and voice let players coach AI teammates to improve outcomes. Uses reinforcement learning agents (built on the SC2LE hack-and-slash maze), adds a language wrapper so commands like heal me or block change behavior in real time, and shows coaching raises success versus pure autonomy. Cites RL scale-ups (LSTM sizes 1024→2048→4096) as precedent, demonstrates playable demos, proposes mobile voice play, Twitch-to-player onboarding, and even selling trained agent weights as digital assets.
AI-written summary · about
chapters (8)
- 0:00Introduction to AI in gaming
- 3:34Evolution of AI bots in games
- 5:40Language as the key to AI control
- 7:47Voice interface and human-AI interaction
- 12:00Empowering players with AI agents
- 15:52Lowering barriers to game entry
- 17:15Human-AI teamwork as the new norm
- 19:13Exploring new dynamics with AI agents
related talks
24mAI as a Teammate: Redefining Agile for Human–AI Collaboration
01h 01mKeynote Panel: Future of AI agents
26mAI Agents and the Future of Human-AI Collaborative Intelligence
56mCan I do AI?
02mDesigning a Video Game AI
26mtalks about being a good AI Coach
01h 01mIntroduction to Practical AI
40mAgentic AI: The Next Evolution of Automation