About Responsible Artificial Intelligence LLC
Responsible Artificial Intelligence LLC (“Responsible AI LLC” or “RAI LLC”) was started by Jiahao Chen in September 2022 to provide his consultancy services on artificial intelligence and its uses in regulated industries such as financial services, employment, education, and healthcare.
What is responsible AI?
Responsible AI is often presented as an umbrella term for AI that incorporates ethics, fairness, explainability, privacy, security, trustworthiness, accountability, and related concerns. Instead, we adopt a simpler, more general definition:
Responsible AI systems are built and operated that take into account the needs of multiple stakeholders
Different stakeholders bring different concerns to the table, which naturally encompass the various concerns that are often treated as a simple checklist to be completed before shipment. However, different people have different needs and desires, which are rarely aligned in practice. Therefore, we believe that the root problem to solve in responsible AI is not simply to address a checklist of concerns, but rather to manage multiple objectives which are often in conflict. A true ethical and responsible AI solution must necesssarily be designed, built and operated with awareness of these conflicting needs, and include strategies for identifying and managing risks and harms that may emerge over its lifecycle. Only by doing so will the risks of using responsible AI solutions be properly managed and its value fully realized by everyone affected.
Responsible AI in practice
Responsible AI LLC brings you the very latest expertise in compliance and risk management for AI/ML systems, even as regulations continue to evolve month to month.
Financial regulatory compliance. We bring extensive experience in building and deploying AI/ML systems in major financial institutions. These environments operate under strict compliance requirements, ranging from:
- data management requirements under international Basel III standards (BCBS 239),
- model risk management standards like OCC 11-12 / SR 11-7 in the US, and upcoming prudential regulations in the UK outlined in CP 6/22,
- and anti-discrimination and fair lending compliance like Regulation B and the Equal Credit Opportunity Act (ECOA) in the US.
Employment regulatory compliance. Governments are increasingly passing new laws imposing requirements on AI systems used in employment decisions. Notably, New York City passed a law on automated employment decision tools (AEDT) which takes effect on January 1, 2023. Proposed Rules for compliance were only published recently on September 23, 2022, which notably contain several requirements that are stricter than the US federal requirements under the Equal Employment Opportunity Act (EEOA) and the many subsequent federal employmnent laws.
Regulatory divergence. Compliance requirements can vary across national, state and local juridisctions, and a compliant solution must take into account such regulatory divergence. Even in something seemingly clear-cut like anti-discrimination laws, different jurisdictions define different taxonomies of protected classes, even down to different taxonomies for categories like race, and have different testing requirements that may or may not account for intersectionality and data availability. Responsible AI LLC provides the very latest expertise on delivering compliant solutions taking into account the real-world complexities around data management, model lifecycle, and operational challenges encounteres in the real world, while continuing to deliver business value.
Responsible AI LLC can work with your data science and MLOps teams to:
- identify and manages multiple conflicting objectives,
- translate ambiguous policies and evolving regulatory standards into practical operational requirements,
- rapid prototyping to explore the design space of potential solutions,
- quantify and manage trade-offs, and
- incorporate all these concerns into the design and operation of responsible AI solutions.
Responsible AI LLC also provides algorithmic auditing services for
- certifying compliance with current and upcoming regulations,
- identify, manage and remediate potential risks of using AI systems, and
- provide expert witness testimony on how compliance needs are addressed in practice.
Responsible AI LLC is fully owned and operated by Jiahao Chen, a respected AI researcher and practitioner who has published many academic papers and has worked at multiple top organizations around the world.
Jiahao’s expertise is widely sought after by many other organizations worldwide. In the past year, these included government agencies like:
and researchers at universities like:
Stay current with responsible AI
Responsible AI is evolving rapidly. Every month, new laws are passed, existing regulations are updated to cover AI/ML systems, and new policy and AI research continues to be published at an exponential rate.
Responsible AI LLC is pleased to offer This month in responsible AI, a newsletter containing brief surveys of the latest developments in academic research, government policy, and community events. Subscribe for free on .