We welcome anyone to contribute to this repository. Please raise issues or pull requests for any missing papers, datasets, or methods. We will update the repository regularly.

• Introduction
• Datasets
  • Single-Vehicle Perception Datasets
  • Motion Forecasting and Cooperative Driving Datasets
  • Simulation Based Datasets
  • Language-Based Datasets
• Methods
  • Image Generation Methods
    • Controllable Generation
    • Decompositional Generation
  • LiDAR Generation Methods
  • Trajectory Generation Methods
  • 3D Occupancy Generation Methods
  • Video-based Scene Generation Methods
  • 3D/4D Generation Methods
  • 3D Scene Editing Methods
  • LLM-based Autonomous Driving Systems
  • MLLM-based Autonomous Driving Systems
• Citation

Note:
For the "Condition" column:
I = Image, T = Text, E = BEV, B = Bounding Boxes/Layout,
D = Depth, C = Camera, M = Maps, A = Driver Action,
O = Optical Flow, J = Trajectory, S = Subject, H = High-level instructions (Command, Goal Point).
Conditions in brackets are optional.

*: Not released as of June.2025.

Note:
In the "Condition" column:
M = Maps, I = Images/Videos, B = 3D Bounding Boxes/Layout, J = Trajectory, T = Text, O = Opacity, C = Camera, A = Driving Action.
* means not presented in the original paper but supported later.
\dagger means reconstruction models with a generative prior.

Here we note their supported operations and output format.

In the condition column, QA stands for question answering, DM for decision making, ED for environment description, SU for scene understanding, and DC for driving context.

In the condition column, VQA stands for visual question answering, SU for scene understanding, DS for driving scene, MVF for multi-view frame, and TC for transportation context.

If you find this repository useful for your research, please consider citing the following paper:

@article{wang2025generative,
    title={Generative AI for Autonomous Driving: Frontiers and Opportunities},
    author={Yuping Wang and Shuo Xing and Cui Can and Renjie Li and Hongyuan Hua and Kexin Tian and Zhaobin Mo and Xiangbo Gao and Keshu Wu and Sulong Zhou and Hengxu You and Juntong Peng and Junge Zhang and Zehao Wang and Rui Song and Mingxuan Yan and Walter Zimmer and Xingcheng Zhou and Peiran Li and Zhaohan Lu and Chia-Ju Chen and Yue Huang and Ryan A. Rossi and Lichao Sun and Hongkai Yu and Zhiwen Fan and Frank Hao Yang and Yuhao Kang and Ross Greer and Chenxi Liu and Eun Hak Lee and Xuan Di and Xinyue Ye and Liu Ren and Alois Knoll and Xiaopeng Li and Shuiwang Ji and Masayoshi Tomizuka and Marco Pavone and Tianbao Yang and Jing Du and Ming-Hsuan Yang and Hua Wei and Ziran Wang and Yang Zhou and Jiachen Li and Zhengzhong Tu},
    year={2025},
    eprint={2505.08854},
    archivePrefix={arXiv},
    primaryClass={cs.CV}
}