Responsible AI Framework for Autonomous Vehicles: Addressing Bias and Fairness Risks
Autonomous Vehicles (AVs) hold immense potential to revolutionize transportation, yet their deployment raises significant concerns regarding safety, security, and ethical considerations. Furthermore, the increased use of automation, edge computing, and Artificial Intelligence (AI), fueled by generat...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2025-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10947002/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1849702113435713536 |
|---|---|
| author | Abhinav Tiwari Hany E. Z. Farag |
| author_facet | Abhinav Tiwari Hany E. Z. Farag |
| author_sort | Abhinav Tiwari |
| collection | DOAJ |
| description | Autonomous Vehicles (AVs) hold immense potential to revolutionize transportation, yet their deployment raises significant concerns regarding safety, security, and ethical considerations. Furthermore, the increased use of automation, edge computing, and Artificial Intelligence (AI), fueled by generative AI technology, has increasingly elevated the risks of AI. Most of the existing research only covers ethical components and lacks the breadth of Responsible AI (RAI), while some have presented components such as justice and solidarity but do not provide an approach or mechanism to implement those in an AI-based system. This paper addresses these gaps by proposing a comprehensive RAI framework for AVs with four key contributions. First, it provides an in-depth analysis of AI risks in AVs, encompassing safety, security, ethical, and legal domains, offering a structured classification to guide developers and regulators. Second, it introduces a holistic RAI framework that spans the AI lifecycle, identifying and mitigating risks at each AV system development and deployment stage. Third, the paper focuses on bias and fairness within AV systems, outlining precise techniques for bias detection and mitigation across data collection, algorithm design, and real-time decision-making. Lastly, the research presents bias removal simulations using publicly available AV datasets, evaluating mitigation strategies such as synthetic data generation and algorithmic fairness analysis. The proposed framework and simulations provide a practical foundation for integrating RAI principles into AV development, ensuring safer, fairer, and more accountable AI-driven mobility solutions. |
| format | Article |
| id | doaj-art-9347d7fc5aa54b0580bd64a963cf8d40 |
| institution | DOAJ |
| issn | 2169-3536 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj-art-9347d7fc5aa54b0580bd64a963cf8d402025-08-20T03:17:44ZengIEEEIEEE Access2169-35362025-01-0113588005882210.1109/ACCESS.2025.355678110947002Responsible AI Framework for Autonomous Vehicles: Addressing Bias and Fairness RisksAbhinav Tiwari0https://orcid.org/0000-0002-6298-5912Hany E. Z. Farag1https://orcid.org/0000-0002-9098-3092Department of Electrical Engineering and Computer Science, York University, Toronto, ON, CanadaDepartment of Electrical Engineering and Computer Science, York University, Toronto, ON, CanadaAutonomous Vehicles (AVs) hold immense potential to revolutionize transportation, yet their deployment raises significant concerns regarding safety, security, and ethical considerations. Furthermore, the increased use of automation, edge computing, and Artificial Intelligence (AI), fueled by generative AI technology, has increasingly elevated the risks of AI. Most of the existing research only covers ethical components and lacks the breadth of Responsible AI (RAI), while some have presented components such as justice and solidarity but do not provide an approach or mechanism to implement those in an AI-based system. This paper addresses these gaps by proposing a comprehensive RAI framework for AVs with four key contributions. First, it provides an in-depth analysis of AI risks in AVs, encompassing safety, security, ethical, and legal domains, offering a structured classification to guide developers and regulators. Second, it introduces a holistic RAI framework that spans the AI lifecycle, identifying and mitigating risks at each AV system development and deployment stage. Third, the paper focuses on bias and fairness within AV systems, outlining precise techniques for bias detection and mitigation across data collection, algorithm design, and real-time decision-making. Lastly, the research presents bias removal simulations using publicly available AV datasets, evaluating mitigation strategies such as synthetic data generation and algorithmic fairness analysis. The proposed framework and simulations provide a practical foundation for integrating RAI principles into AV development, ensuring safer, fairer, and more accountable AI-driven mobility solutions.https://ieeexplore.ieee.org/document/10947002/Responsible artificial intelligenceethical artificial intelligenceautonomous electric vehiclesgenerative artificial intelligencesynthetic data |
| spellingShingle | Abhinav Tiwari Hany E. Z. Farag Responsible AI Framework for Autonomous Vehicles: Addressing Bias and Fairness Risks IEEE Access Responsible artificial intelligence ethical artificial intelligence autonomous electric vehicles generative artificial intelligence synthetic data |
| title | Responsible AI Framework for Autonomous Vehicles: Addressing Bias and Fairness Risks |
| title_full | Responsible AI Framework for Autonomous Vehicles: Addressing Bias and Fairness Risks |
| title_fullStr | Responsible AI Framework for Autonomous Vehicles: Addressing Bias and Fairness Risks |
| title_full_unstemmed | Responsible AI Framework for Autonomous Vehicles: Addressing Bias and Fairness Risks |
| title_short | Responsible AI Framework for Autonomous Vehicles: Addressing Bias and Fairness Risks |
| title_sort | responsible ai framework for autonomous vehicles addressing bias and fairness risks |
| topic | Responsible artificial intelligence ethical artificial intelligence autonomous electric vehicles generative artificial intelligence synthetic data |
| url | https://ieeexplore.ieee.org/document/10947002/ |
| work_keys_str_mv | AT abhinavtiwari responsibleaiframeworkforautonomousvehiclesaddressingbiasandfairnessrisks AT hanyezfarag responsibleaiframeworkforautonomousvehiclesaddressingbiasandfairnessrisks |