Booz Allen’s purpose and values have served as our “North Star” when navigating technological transformation for 110 years. Our unwavering commitment to these powerful truths is the foundation of our AI practice and delivery. We hold them to be immutable—unchanged by past (or future) technology waves.
We have codified these Purpose and Values as a series of Guiding Principles for designing, developing, deploying, and using AI systems. These Principles serve as the foundation for Booz Allen’s AI delivery framework.
We believe that alignment of AI with these purpose and values helps ensure we can meet our commitment to deliver AI at the scale and speed to mission – ensuring accuracy, adoption, as a function comprised of ethics, accountability, and safety considerations as elaborated in our Guiding Principles below. These Principles must inform every AI application, but ensuring successful AI necessitates balancing the unique risks, challenges, and context of each specific use case.
Our AI Guiding Principles solidify our commitment to AI practices.
Booz Allen commits to our values of ferocious integrity, unflinching courage, passionate service, champion’s heart, and collective ingenuity when driving AI innovation. We adhere to these values when identifying opportunities where the public benefits gained from the design, development, and use of AI significantly outweigh the risk—and where the problem identified can be solved in a manner that meaningfully and effectively manages any such risk. We bring judgment and care to our AI activities with the fundamental goal of doing right by our clients, our people, and the American public. We believe AI systems should be trustworthy and mission enhancing.
Booz Allen designs, develops, and implements AI solutions with ferocious integrity by ensuring clarity of who is responsible for all phases and aspects of AI systems in order to provide clear accountability. We seek to continuously incorporate evolving transparency practices and tools in developing AI systems, strive to confirm that the ongoing use and outputs of such tools remain as intended, and assess the same practices when evaluating the technologies we procure and use in our own operations. We believe that transparency should increase proportionately and proactively as the consequences and risks associated with AI systems increase.
Booz Allen exercises unflinching courage in design, development, and deployment of AI systems to ensure they do not unintentionally cause physical or digital harm. We design with a champion’s heart, craving to be the best, and seeking to ensure the safety, security, and resiliency of AI applications and the protection of all data contained therein against threats of systematic vulnerabilities, adversarial manipulation, or other malicious exploitation.
Booz Allen conducts robust analysis so that AI systems remain consistent with the use cases for which that system was trained and that ongoing use remains accurate, reliable, and effective. Booz Allen is committed to the use and implementation of AI technology that is robust against high-paced technological enhancements and advanced adversarial capabilities. We bring collective ingenuity to continually invest in improving our approaches and tools, harvest lessons learned from delivery, and continually improve the resiliency of our AI solutions.
Booz Allen seeks to minimize harmful bias throughout its design, development, and use of AI systems—holding ourselves and others accountable to do the right thing. We act with empathy and harness talent with valuable and unique insights, addressing harmful systematic and statistical bias, considering vulnerable populations, and employing appropriate checks from both internal and external stakeholders from affected groups to help mitigate bias and distortion in the application of AI. We leverage our collective ingenuity throughout the AI lifecycle to help ensure that AI advances dignity and fairness—and to anticipate and mitigate the impacts of harmful bias and be transparent about such shortcomings if they exist.
Booz Allen is committed to the lawful and ethical development and use of AI tools and systems that embody respect for passionate service to protect societal values and individuals’ rights and interests, and adhere to all other applicable laws and policies. When developing and deploying AI systems, we work to safeguard individual, group, and entity privacy, as well as civil rights, civil liberties, democratic values, human rights, equal access, intellectual property, and the wellbeing of workers, society, and the planet. This includes applying human intervention as an alternative to AI where appropriate to protect the public from harmful impacts.
Exercising unflinching courage, Booz Allen holds itself, Booz Allen people, its clients, its suppliers, and other stakeholders accountable and thoughtfully considers each use case to ensure that all relevant individuals understand how data is being used and how each AI system makes decisions. We are committed to reviewing each use case to ensure that stakeholders possess an appropriate understanding of technology, development processes, and operational methods relevant to their role and the specific application. We are focused on ensuring our AI systems have clear and auditable methodologies, data sources, and design and procedure documentation.
Booz Allen actively reviews, assesses, and manages risk in deployed AI systems within the scope of its control and contractual responsibility and encourages its clients to adopt similar practices as methods, contexts, risks, and needs evolve over time. Booz Allen is committed to maintaining human governance in its own systems where appropriate and to encouraging the same for client systems. Booz Allen regularly tests and takes steps to identify, analyze, and mitigate risks throughout the development lifecycle to maintain adherence to these principles, and develops mechanisms for the periodic assessment, review, and decommissioning (if needed) of systems within its control to ensure continued adherence.
