AI Games are immersive and challenging experiences that incorporate generative AI technology. The resulting NPCs can exhibit more human-like behavior, leading to more dynamic and freeform gameplay experiences. Players also often troll and prank these AIs, further increasing immersion in the game world. Resource :

AI in gaming has been a long-term area of research. One of the earliest examples was the computerized game Nim, created in 1951 with an advanced Ferranti Mark 1 machine. This game was able to beat highly skilled human opponents and was the first example of computerized AI.

Real-Time AI for In-Game Commentary

Contemporary AI technologies such as RL and LLMs are applied to create more realistic game AI systems. These are integrated into traditional game AI architectures, making them more believable and interactive.

A key challenge with game AI is to make it ‘hunt’ its enemies rather than simply reacting to them in a black or white manner. Hunting enables the AI to take a more measured approach to an enemy, such as moving into flanks or engaging in a more evasive combat posture, allowing it to better assess the threat and decide on an appropriate response.

Game developers face tight development cycles, which makes it hard to test and understand cutting-edge game AI technology. By using AI, it is possible to automate tedious tasks and make complex decisions with much greater efficiency and accuracy than a human developer. This frees up more time for the creative ideas that make games so compelling.