What do you do with an AI conscientious objector?

Recently, I was confronted with a novel situation regarding a conscientious objector to AI. While working to formulate goals for our AI strategic initiatives, a teammate informed me they would refuse to use it as part of their role. Digital transformation is difficult to achieve on its own, let alone when members of your own team refuse to use new tools to get there. However, this situation is occurring across companies worldwide, as some employees decline to be onboarded with AI due to environmental or ethical concerns.

The conscientious objector on my team was largely concerned about future job displacement. I assured them that I didn’t require them to use something with which they disagree. However, I intend to hold that person to the same benchmarks as everyone else, and those benchmarks now assume the tool and its productivity gain exist. This may create a difficult situation for the employee as we adopt AI technology, but as a leader, I’m responsible for my teams’ performance and objectives.

AI has a PR problem

A healthy dose of skepticism always comes with the territory of any major technological shift. But as teams across industries race to embrace AI, consistent criticisms seem to be cutting through. My employee voiced most of them in our discussions.

The first was the biggest long-term fear: job displacement. The person refused to use the tool because they refused to “train their replacement”. The second was trust (or lack thereof). Their opinion is that AI is a machine-led tool that not only hallucinates but also can make the human in the loop increasingly lazy. The person stated that if the output is automatically presumed correct, we create a beginning-to-end resource with too much power and authority in decision-making. Third, there’s the moral objection (one shared by my wife) that data centers are springing up across the country, and small communities are pushing back against the strain on water and power.

This employee isn’t alone in their concerns. A survey in May by CNBC and SurveyMonkey found that nearly two-thirds (65%) of workers have at some point avoided using AI because of similar concerns, including that the technology can be inaccurate, a detractor to creative thinking and an eliminator of entry-level roles. Some experts perceive the discomfort in almost existential terms as an “attack on humanness.”

Others go even further than declining to use it. Nearly one in three (31%) employees report intentionally derailing their company’s automation roadmap by tampering with performance metrics to make it appear AI is underperforming and intentionally generating low-quality outputs. Interestingly, the number engaging in such practices jumps to 41% for millennial and Gen Z employees. This is a surprising statistic given younger employees tend to be more excited about technology adoption and innovation.

To their credit, my objector sat firmly on the honest side of the line. They weren’t breaking the machine but telling me, openly, that they didn’t want to use it.

Balancing ethical objection and technological progress

My initial response to this stance was to lead with honesty. I shared that AI is a tool, and your job is to get from A to B, while being free to choose the route there. At the same time, however, there are new business realities to keep in mind, particularly when generating specific goals about AI adoption.

At Paessler, like a lot of companies, we’re now setting internal AI benchmarks and moving a meaningful share of manual workloads onto these tools. Once that happens, declining it is no longer choosing a different path to the same result, but is actively stepping outside of something the whole team is now measured against. For example, other teams are using agents to automate routine tasks, share knowledge and plan work. If these cross-departmental efforts are rejected, there’s a risk that the employee will not be able to keep up with company-wide innovations. None of this is a personal criticism or a verdict on someone’s values, but the benchmark itself has shifted considerably in the last year.

There are also negative effects for leaders to consider as they navigate similar situations. If one person’s opt-out is tolerated indefinitely, others reasonably ask why they can’t do the same. Before long, there is a tangible risk that the team’s targets will slip, calling the leader’s tactics into question.

I suggested the employee focus less on the automation risk and more on the augmentation benefits. AI is a tool that can improve both your efficiency today and expand what’s possible tomorrow. For what it’s worth, I truly believe that AI job losses are much more likely to stem from a lack of retraining than from the technology itself. The employee who isn’t trained in AI tools is at much greater risk of losing their role than someone who can effectively supplement their work with the technology.

Encouragement can only go so far, though, when considering how AI has changed the landscape. Nearly every technology company is moving this way and, while it isn’t the posture I’d advocate, roughly 60% of executives say they’re considering laying off employees who won’t adopt AI. Of course, a hardline stance isn’t very motivating, so there needs to be a strategic bridge between “I respect your view” and “this is now a performance issue”. Our answer was to elect an AI ambassador on each team. This ambassador acts as both peer and champion who helps a hesitant colleague set up first use cases, find the value and establish the necessary KPIs. This works because a colleague allaying fears lands differently than a manager issuing mandates.

As a leader, I also couldn’t move forward without addressing trust. I agreed that we can’t just leave the tool to its own devices, but that demands best practices rather than abstention. For me, quality controls offer employee agency in this evolution. If your name is on the process, you own it, including anything the AI got wrong. Due diligence is more important than ever and it’s still up to the team to follow the links, check the conclusions and cut what doesn’t hold up. This is par for the course for any new tool, and I’ve previously caught AI misjudging a source and dropped it accordingly in my own work. Additionally, executives are monitoring output and intent, and they won’t simply accept AI slop. Checks and balances help build that trust and ensure the machine isn’t calling the shots.

Encourage agency, adopt consciously, educate along the way

AI is a massive capability jump and one that can be equal parts exciting and frightening. I think we’ve all had moments of uneasiness with this technology. For example, I remember modeling use cases and encouraging the security team in the early adoption phase. I watched one of my teammates start out with reservations, then be absolutely blown away by what he could suddenly do. Work that typically took weeks got done in hours or days, such as CMMC internal maturity assessments, penetration testing exercises and policy reviews. Once we entertained that curiosity and had team members thinking about how they could incorporate it, the results were remarkable.

For me, this is a reminder of how leaders need to approach this moment. Frame the technology as something that delivers a net benefit, removing rote work and freeing up time for higher-impact tasks. But, crucially, remind your team that the work still depends on them for oversight and guardrails. Where the objection is environmental, show some flexibility by encouraging leaner usage rather than a blanket refusal, because you can manage the footprint, but you can’t use it to opt out entirely.

Amidst the hype, many of us have forgotten that AI is a tool. It’s impressive, flawed and useful, but ultimately a tool. Refusing it outright is refusing a transformative technology, and putting your head in the sand isn’t a solution. At the same time, there are legitimate issues that employers and employees need to work out together.

This debate takes me back to the introduction of cloud computing and similar fears in the IT industry. Concerns such as the elimination of data-center jobs and the gutting of infrastructure teams were common. Long-term, though, it didn’t play out that cleanly, as there became a sudden need for cloud IT and security engineers. The cloud was soon understood as a different kind of data center requiring different skills, and the ones who refused were left behind. The best way forward was to continually adopt and adapt, and the same is true for any type of technical innovation.

Ultimately, for many of us working at technology companies, it’s hard to be a technologist when you refuse innovative technology.