Responsible AI Guidelines 2024 Table of Contents Introduction 3 About this Handbook 4 Our AI Vision Our AI Mission Our Four Principles of Responsible AI 8 Human-Centered AI Assessments 12 Empower Personas in the AI Lifecycle Responsible/Human-Centered AI Assessments Human Collaboration, Feedback, and Oversight Human Collaboration, Feedback, and Oversight Assessment Inclusive 20 Accessibility and Diversity Evaluation for Inclusivity & Bias Assessment Transparent 24 Prioritize transparency, traceability, and contestability Resources Transparency, Interpretability & Contestability Assessment Accountable 28 Accountability Statement Mitigate Undesired Behaviors and Outcomes Checklist Understanding the Potential Benefits and Harms of AI Exploring impact, risks, opportunities, and trade-offs behind each AI Solution Set Realistic Expectations Realistic Expectations Assessment Continuous Risk Monitoring and Mitigation Evaluate Early and Often Evaluation Assessment Data Considerations Social and Environmental Impact Social and Environmental Impact Assessment The ServiceNow Responsibility Closing Statement 43 Introduction In the realm of AI, user experience design and research are not merely steps in the process—they are the keystones of innovation. By grounding AI development in the richness of human diversity and real-world complexity, we ensure technology serves us with relevance and humanity. Human-centered AI is the development and deployment of AI systems that prioritize human needs, values, and well-being. This approach involves designing AI systems that are transparent, responsible, and auditable. Ultimately, human-centered AI aims to create technology that serves humanity and aligns with our aspirations, goals, and values. This guidebook serves as a primer for anyone using or working with ServiceNow’s AI technology in their careers or businesses. Readers will learn practical guidance and valuable information on ways to navigate using AI responsibly. It underscores the need for AI systems that prioritize human needs, values, and well-being, advocating for technology that respects and enhances the human experience. ServiceNow’s Code of Ethics ServiceNow’s Human Rights Statement ServiceNow Trust and Compliance Center 3 About this Handbook Practical guidance and decision-making tools People who use ServiceNow’s AI technology To create a better world of work through the appropriate and responsible use of AI Best practices advocate for regular assessments What For Why When This initiative represents our continued dedication to offer practical guidance and consistent decision- making tools to all individuals involved in building, configuring, and consuming ServiceNow’s AI-driven products and features. Familiarizing yourself with the content in this handbook is strongly encouraged if you are contributing to an AI-driven product experience or considering incorporating AI into your work. Resource: NIST AI RISK MANAGEMENT FRAMEWORK LINK 4 As a note, we are using “AI” as a large bucket term that falls across the spectrum and includes various forms of AI including machine learning, generative AI, and LLMs (large language models). While this latest version of the Responsible AI Handbook is designed to provide immediate value, we acknowledge that there is ongoing work to be done. This is not an exhaustive set of guidelines and principles and much of it is aspirational, because there are aspects of technology and building responsible AI that is ongoing, in research and development. As our products evolve, technology improves, and the market shifts, we will continue to iterate, and release updated revisions of this handbook. We have also included worksheets with key questions for each section to help guide meaningful discussions and decision-making within your company and teams. 5 Our AI vision As we look toward the future, we envision a world where ServiceNow products leverage AI in meaningful ways to drive better outcomes for our customers. We envision a seamless integration of AI across our ServiceNow suite, with tools that not only streamline the creation, configuration, and monitoring of AI services but also democratize AI utility, making it accessible irrespective of expertise in data science. We are building a future that is conscientious about AI ethics, fostering trust, and embedding our corporate ethos into our AI-driven solutions. Our journey ahead is focused with purposeful innovation and a commitment to harness the transformative power of AI to elevate human collaboration and creativity in the workplace and enable individuals to work more effectively. 6 Our AI mission At ServiceNow, our mission is to use the transformative power of AI within workflows,
driving enhanced performance and world-class experiences. We provide our valued customers with powerful and reliable AI tools that empower innovation and unparalleled solutions. We reimagine workflows and augment existing ones with AI- based efficiency gains. We focus on the people, use cases, and opportunities that benefit most from AI. With you in mind, we've established a set of comprehensive pillars that steer our AI technology development. These pillars are founded on the industry's best practices and ensure the creation of trustworthy and responsible AI for you and your customers. Each of these pillars are meant to guide you through the AI development cycle, from initial creation by our team to practical application by yours, enhancing the way your end - users and customers interact with our services. 7 Our Four Principles of Responsible AI Our AI mission Your text goes her Add bullets by making a new lin Indent the bullets by pressing tab Our Four Principles of Responsible AI Four Pr nciples of Responsible AI Trust is earned in drops but can be lost in buckets. ServiceNow recognizes the importance of responsible AI development.
When developing new products and services for our customers, we center our approach around the following principles: 1 Human-Centered 2 Inclusive 3 Transparent 4 Accountable 9 1. Human-Centered 2. Inclusive We enable our customers to choose where and when they leverage AI in our products by clearly identifying when AI is being used, keeping humans in the loop. Our in-product messaging and product documentation clearly identifies when AI is used. We provide our customers with this guidance, so they understand how to leverage and deploy our AI in a responsible manner. We believe in the power of technology to reduce complexity and make the world a more equitable place. Our AI teams are broadly diverse and test our AI systems with datasets that reflect the diversity of our customers and our end users. Used responsibly, we believe AI can improve how all humans interact with technology to achieve their goals and aspirations. 10