Skip to main content
01
Worker safety in the age of AI
Article by
Dr. Elizabeth Adams
Chief Executive Officer, EMA Advisory Services
5 FEB 2025 | 5 min read
As workplaces evolve, so too should the partnership between workers and employers.
How many employees today can say they have a partnership with their employers? How many employees can say they feel they work in a safe environment where they can share their ideas, hopes, and fears when it comes to AI design and development? Answers to these two questions might just suggest the social contract between employers and workers may be at risk.
The signs of working confidence eroding are evident across headlines, social media platforms, and the research that I and countless others conduct. A deepening desire for next-generation artificial intelligence is a curious one and yet is on a fast course headed for collision if we are unable to address societal notions about the longevity of certain jobs, the threat of sunsetting skills, and general unease about the value of the employee stakeholder in increasingly automated workplaces.
Whether AI anxieties are sustained in the long-term remains to be seen. Yet, here might be a clue: a significant percentage of workers feel deep concern, as is evidenced by the Pew Research Center, the American Psychological Association, the Edelman Trust Barometer, and others. Addressing employee safety concerns early, doesn't have to end up being a costly roadblock for enterprises seeking to adopt and scale AI, while also seeing a return on their investments. There are frameworks that can be adopted if an organization is interested in strengthening its partnership with its employees.
To that end, many organizations around the world have set out to establish a vision for AI which includes cascading responsible AI governance practices across their operations, a necessary first step. But what comes next? I suggest organizations should prioritize creating an environment where employees feel safe to express their optimism and fears about AI. By engaging employee stakeholder practices and adopting a Worker Safety Statement, organizations can foster a collaborative atmosphere that supports both the workforce and the successful integration of AI. Engaging employees as key constituents is a critical element in an organization's AI adoption journey.
The terminology around worker safety evokes images of factory floors, construction sites and other professional spaces more often associated with physical labor. But our idea about worker safety — and what it means — needs to evolve, just as AI is forcing so much of the market to evolve.
Trust in AI begins with safety. Employees who feel safe — psychologically, professionally, and ethically — are more likely to trust their organization's AI initiatives and more likely to be enthusiastic about contributing to the organization's success. Safety goes beyond avoiding physical harm; it encompasses the assurance that employees can speak up without fear of retaliation, contribute without exploitation, and challenge AI practices through healthy discourse. When workers see their perspectives valued, they not only trust AI systems more but also become active participants in shaping them.
The landscape of the modern workplace is fast evolving as organizations figure out ways to implement emerging technologies like AI. And it is against this backdrop that tech leaders have every responsibility to ensure the safety and insertion of their employee's perspectives. Worker safety, both in terms of physical environments and ethical participation, is a cornerstone of responsible AI practices. Without a commitment to this principle, organizations risk not only eroding trust but also compromising the very systems they aim to enhance.
Creating a safe environment for employees to express their concerns and ideas about AI is crucial for building trust and fostering innovation in the workplace.
Defining worker safety
Through my own research — which most recently has included conducting 51 interviews across 15 companies and then sifting through more than 3,000 hours of transcripts — I've managed to document one vital recurring theme. Many employees and their colleagues feel excluded from critical conversations about AI. They recognize their role as stakeholders and as having a vested interest in the success of their organizations. They also understand their role in representing customers and clients. Yet they lack an organizational path to voice concerns or contribute insights. This absence of a documented and practiced path weakens trust, stalls innovation, and impedes the organization's ability to develop AI responsibly.
In my conversations, I've asked the question, "How might you participate in Responsible AI practices if asked?" One poignant example involved a customer care employee unable to answer questions of a customer regarding an AI chat bot in operation. The employee suggested the oversight in gaining their perspective led to frustration and the inability to assist a customer at a critical time in the customer's journey. To address this, companies must equip employees with training, tools and engagement in testing AI systems they will be impacted by, ensuring they can confidently engage with AI and represent its functionality to external stakeholders. Further, workers should have a reliable and safe path for voicing concerns.
Taking the proactive step of communicating to employees that their insights and concerns are seen and taken seriously is a critical step, even if it might sound unconventional. Why is it important? Because by centering the employee in how an organization shapes the future of AI at your organization, it telegraphs to workers that leadership values them enough to publish an official statement.
The adoption of AI is not seamless. It is nuanced. Just think of how we learned to embrace the adoption of cybersecurity initiatives. It's similar. Mistakes will be made, glass will be broken, repairs will have to take place, but enthusiasm can be a parallel experience. A statement on worker safety at least communicates to the workforce that their organization is trying to do things right.
So what does right look like?
For starters, there should be a path for employees to address their claims, a digital platform for workers to submit their ideas, observations and concerns. Consider something akin to an employee led townhall with a formal representative — a safe convening spot and a methodology for participation.
Taking the first step
As with many aspects of deploying AI across an organization, it is particularly important to have buy-in not just from the top of an organization's leadership, but across the top. Virtually every role within the C-Suite should have an interest in worker safety.
Let's think about how leaders can map employee involvement back to the frame of what every worker safety statement should convey. First and foremost is transparency around the vision for AI and the use case for the adoption of AI tools and how employees can contribute to the organization's success. Clear communication about how AI systems are developed, governed, and used is critical and why your organization has chosen this case. Next is engaging employees to design a pathway for feedback, which can include introducing a digital platform, formal groups for reviewing policies and procedures, creating cross-functional testing teams to engage with AI applications before implementation. Consider that a combined responsibility of Corporate Affairs, the CIO and CHRO.
The worker safety statement should also promote access to training programs to educate employees about AI systems and equip them to engage meaningfully — something the CHRO and CTO are positioned to address. And finally, there is the criticality of the statement showcasing to employees a protection from retaliation. This element calls for partnership between the CHRO, Chief General Counsel, and possibly employee representatives and it should ensure a safe environment where employees can speak up with enthusiasm.
Consider the statement an evolving artifact that brings a sense of organizational pride to an environment. This statement does not need to be perfect at the start. In fact, acknowledging its imperfections signals a willingness to grow and adapt, something that I would consider a critical element of responsible AI.
Short of these steps, there must be acknowledgement that there is a cost of doing nothing. What we want to minimize is inviting unnecessary risks, including reputational damage should an AI misstep undermine public trust. Employees, sit closes to the actual work of making AI "happen" in an organization, and can quickly inform your processes when they are off track.
The truth is that safe and responsible AI isn't a luxury, it's an imperative.
Centering the employee
In my research I learned of the value of including in AI design and development discussions about AI's societal impact from employees whose use of AI outside of the organization is growing. Based on their growing adoption and expertise of AI use at home or in their private time, employees are uniquely positioned to link their company's goals with their own responsibilities when afforded a path to do so. By centering employees, organizations gain access to lived experiences, cultural intelligence, and frontline insights. This is particularly crucial for employees, whose perspectives can help catch blind spots in AI design and governance. Employees are energized by AI and see themselves as active stakeholders whose insights can drive ethical innovation.
Creating a Worker Safety Statement is a bold and optimistic action, signaling a commitment to fostering a safe, and innovative environment. Organizations can demonstrate their appreciation for employees by engaging them in shaping the future of work. And most importantly, they lay the groundwork for trust, a perfect partnership that can be felt, between workers and organizations, between AI and the world.