AI systems that have evolved to adapt to new environments often fall under the umbrella of reinforcement learning, where agents learn by interacting with their environment and receiving feedback in the form of rewards or penalties. Some notable examples include:

1. **AlphaGo and AlphaZero**: DeepMind's AlphaGo initially learned to play the board game Go by analyzing millions of human-played games. However, AlphaGo's successor, AlphaZero, learned to play Go, chess, and shogi solely through self-play, adapting its strategies to different games and environments.

2. **OpenAI Five**: OpenAI's team of AI agents, OpenAI Five, learned to play the complex video game Dota 2 at a high level by playing millions of games against itself, continually adapting and improving its strategies over time.

3. **Robotics**: AI-powered robots, such as those developed by Boston Dynamics, learn to navigate and interact with their surroundings through reinforcement learning and other techniques. These robots adapt to new environments, terrains, and tasks, such as moving boxes in a warehouse or traversing uneven terrain.

4. **Autonomous Vehicles**: Self-driving cars use AI algorithms to navigate through various environments, including city streets, highways, and rural roads. These systems continuously learn from their experiences to improve their ability to handle diverse driving conditions.

These examples demonstrate how AI systems can evolve and adapt to different environments through learning and interaction.