3 ways ServiceNow’s people team embraces responsible and ethical AI

AI is rapidly transforming the workplace. When it’s used appropriately, organizations reap the benefits of increased productivity, efficiency, and cost savings.

Boosted capacity can fuel more innovation and drive more significant and better outcomes. However, AI has the potential for risk if used carelessly and irresponsibly. Unethical AI in HR is especially problematic and requires rigorous oversight and governance. This is because of the ever-changing regulatory landscape.

At ServiceNow, we’re committed to ensuring our AI respects human values, prioritizes transparency and accountability, and fosters equality and fairness. Here’s how our global people team helps safeguard responsible and ethical AI while remaining compliant and maintaining a human approach.

1. Establishing an HR AI governance model

An AI governance framework sets up guardrails to help oversee policies, procedures, and ethical considerations for the development, deployment, and maintenance of AI systems. An effective AI governance model is multifaceted and comprises various focus areas, including assessing and mitigating risks, creating standards, assessing bias and fairness, and monitoring and reporting.

Responsible AI requires deliberate, cross-functional collaboration between teams. Legal, IT, compliance, data governance, data privacy, and employment law teams should all play prominent roles.

In 2024, the ServiceNow global people team established an HR-focused AI governance model to help ensure ethical and responsible AI within the team and HR technical products. This supports the team in meeting all compliance and regulatory requirements.

Our model is especially important given the rapidly changing AI regulatory landscape and the introduction of numerous laws worldwide, such as the EU AI Act.

Our HR AI governance model is led by a steering committee comprised of crucial global people leaders and experts in AI, legal, and IT matters. This committee aligns our AI product roadmap and existing technology with the business to select and use ethical AI solutions.

In addition, the model helps us assess and identify risks, measure the success of our AI programs, and enable employees to use the AI tools we invest in.

2. Ensuring ethical AI use in HR

Ensuring fairness and ethical AI is a priority for ServiceNow’s global people team. Our HR AI use cases go through a multistep scoping and risk assessment process before we implement (or build) an AI-powered system or application.

We don’t buy or build AI for the sake of having AI. We scrutinize each use case to develop a solid understanding of the benefits and implications from all angles. Before we deploy any AI tool, we conduct a full risk assessment to uncover any ethical concerns and potential misuses that could harm our employees.

Not all AI used in an employment setting is high risk, but some use cases require extra scrutiny. At ServiceNow, we don’t use AI-driven automated decision-making in employment matters. We require a human in the loop, robust bias testing and reporting, and ongoing monitoring. This allows us to safely invest in new AI technologies that bring value to the organization while protecting our employees.

Bias in AI models is a known problem. Many commercially available large language models and prediction engines were built on data sets with biases hidden within. Custom AI development is also prone to hidden biases, especially in HR.

Testing for bias and fairness is a standard at our company. After we thoroughly evaluate our AI use cases on the front end, the AI undergoes checks and balances during development and deployment. As a precaution, we employ rigorous training and testing standards to reduce the chance of bias creeping into our AI models.

3. Giving HR leaders visibility

Organizations need visibility into AI systems and applications to manage those investments effectively. Without oversight of the inner workings of AI models, it can be difficult to monitor AI behavior, maintain compliance with employment and AI laws and regulations, make technology decisions, and address the ever-changing needs of the business.

AI model monitoring is often focused solely on the technical aspects of AI and minimally considers ethical and responsible use. At ServiceNow, we built AI Control Tower, a streamlined AI asset management tool with advanced governing and monitoring capabilities. It enables us to see our AI inventory, track the performance of our AI models, incorporate risk and security assessments, and audit value monitoring and realization.

By prioritizing ethical AI within our business, we realized that HR has a vital role to help ensure responsible AI. We care deeply about employees' experiences and are obligated to provide (and measure) fairness and the absence of bias in the workplace.

Prioritizing AI governance for the future

We’re at a pivotal point in human history where AI (when used ethically and responsibly) is becoming a massive catalyst for the growth and development of the workforce. Automating mundane tasks gives workers the capacity to explore new ideas and foster innovation.

At ServiceNow, we embrace this new world but also see the current limitations (and risks) of rapidly advancing technology. We plan to move fast and continue to innovate and automate while using AI safely, ethically, and responsibly.

Find out more about how ServiceNow helps organizations embrace AI in HR Service Delivery.