We are committed to upholding strong ethics as part of our core business approach—not just because we believe it gives us a competitive advantage, but because it is the right thing to do.
As technology evolves, new opportunities arise; with them, however, can come the potential for unintended effects. These innovative new tools must also be used in a way that is ethical, secure, and compliant.
ADP has adopted a rigorous set of principles and processes to govern its use of artificial intelligence (AI), machine learning (ML) and other newer technologies. We continuously work to safeguard our trusted dataset by building in rigorous standards to deliver unbiased, independent and objective insight into the workforce.
We understand the great opportunity that new technologies like generative AI provide as we design and develop innovative solutions that address the critical needs of our clients and associates. As the power of generative AI grows, we’re remaining laser focused on the human experience, providing easy, smart and human technology to help HR practitioners, managers and employees make informed decisions more efficiently and to achieve their goals confidently.
As ADP works to provide clients the benefit of generative AI, we continue adopting use cases while ensuring a strong compliance framework. We have developed a compliance-by-design process that establishes key guardrails around transparency, bias, and accuracy of our generative AI tools. This process also integrates privacy, security and intellectual property principles into our ideation and development phases for generative AI use cases and to monitor the use of our generative AI tools. This process involves interdisciplinary teams across our organization to ensure we approach this new technology in an ethical, secure and compliant way that is beneficial to our clients.
ADP believes that human oversight is essential to the reliable operation of AI and ML models and making proper use of their results. Our solutions provide recommendations to human decision-makers, which they can then decide how to act upon.
We strive to develop ML models that are explainable and direct, with clear purposes. Our ML models are designed with understanding as a key attribute, measured against an expressed desired outcome. We test and evaluate our ML models accordingly, adjusting as needed to maintain accuracy in line with the models’ purposes. We provide our clients with information about how our ML models operate, their proper use, and their limitations, so that clients can implement those models in accordance with their design and purpose, operate them effectively, and use their outputs as intended.
ADP’s approach to AI emphasizes the isolation of unintended bias. We are vigilant not to reproduce bias in any AI-enabled product or service. Even when accounting for potential unintentional bias in source data, coding, or use of an AI-enabled product or service, there can be unexpected or unforeseen bias that come into play. ADP’s goal is to continually strive to identify new and unexpected sources of bias and then refresh and enhance the design of our client offerings to address them.
ADP’s MLOps program is a set of processes and automation for managing models, data and code to improve performance, stability and long-term efficiency in ML systems. AI based applications begin with model creation using specially prepared data, which then trains the model. Once the model is developed and ready to deploy in an application, it is registered in our MLOps systems, to enable monitoring of the model’s performance and to obtain feedback which then goes into model improvement. We monitor models for fairness, performance, and drift, so that they are operating as intended. Our governance process includes a multifaceted review by a cross-functional working group bringing together business, technical, legal, security, and compliance experts across the company.
We have an active AI & Data Ethics Council, comprised of both industry leaders and ADP experts across our business, which meets on a regular cadence and reviews our design principles. The Council advises on emerging industry trends and concerns and provides guidance with respect to the principles ADP should follow while developing products, systems and applications that involve AI and data.
We are committed to having diverse teams design and develop our ML models, to ensure a wide variety of perspectives and experience are considered. After all, ML models impact humans, and human experience should inform that impact. In addition, we support skills development to accelerate the growth of a diverse workforce that can develop and deploy the AI solutions of the future.
See how ADP’s AI principles are embedded in our AI-driven tools to assist recruiters.
At ADP, security is integral to our products, our business processes and our infrastructure.
Our associates reflect ADP's commitment to privacy every day in their actions and commitments.
