- This article was originally published in Dutch on 25 September 2025. It was written by Philip Verhaeghe of the NextConomy .
Whilst companies worldwide are investing massively in artificial intelligence, as many as 40% of Gen Z workers are sabotaging these initiatives. This was revealed by Professor Stefano Puntoni of The Wharton School during a recent international online conference organised by ‘Reshaping Work’.
Since late 2022, organisations worldwide have been experimenting with various generative AI technologies. Puntoni, co-director of Wharton Human-AI Research, presented research that exposes an alarming gap between “optimistic executives” and “anxious workers”.
“We are at a pivotal moment,” said Puntoni. “Algorithmic systems now mediate a wide array of work practices. They blur the boundary between humans and machines. But you can have the most sophisticated AI technology — if half of your workers resist, your strategy will fail with certainty.”
The Paradox of ‘Optimism Versus Anxiety’
Three years of longitudinal research amongst senior decision-makers at major American companies shows a remarkable evolution. In 2023, 80% believed that AI would enhance employee skills whilst 75% believed that AI would replace employees. This nearly equal distribution has shifted dramatically: in 2024, belief in enhancement rose to 90%, whilst concerns about replacement fell to 72%.
“Executives are developing a more nuanced understanding of AI’s capabilities and limitations through practical experience,” explains Puntoni. “They see AI as a productivity catalyst and opportunity for growth. But their top-down optimism clashes dramatically with the reality on the ground.”
That reality is concerning: massive resistance and unrest, strikes where AI is central, and systematic undermining of business strategies. The consequence is that billion-euro investments in AI technology are failing due to human resistance.
Puntoni’s team applied Deci and Ryan’s Self-Determination Theory to understand why so many workers experience AI as an existential threat. He illustrated this with striking examples. Professionals do not resist technology per se, but they do resist technology that renders their expertise irrelevant. And whilst algorithmic management can optimise individual productivity, it simultaneously destroys the social infrastructure that enables innovation and knowledge sharing.
“Billion-euro investments in AI technology are failing due to human resistance.”
Recommendations for smooth collaboration
Based on his research findings, Puntoni formulated five interesting implementation principles for business leaders:
- Conduct systematic assessments to identify which workers and roles experience the greatest identity threats. Data analysts respond differently to AI tools than customer service representatives. Therefore, adapt your approach accordingly.
- Build AI systems that require human judgement, creativity, and oversight. Avoid ‘black box’ solutions that reduce people to button-pushers. Make the human role central, not peripheral.
- Budget 30 to 40% of AI implementation costs for training, reskilling, and psychological support. This is not overhead but mission-critical infrastructure.
- Be extremely cautious with fully algorithmic management. Human relationships and social connections must remain a design priority, even with AI-supported decision-making.
Workers must be able to experiment safely, make mistakes, and voice concerns. Organisations that punish AI scepticism drive resistance underground where it becomes more destructive.
He concluded with several essential core recommendations for sustainable “human-AI partnership”. For organisations that neglect the psychological dimension of AI adoption risk not only the failure of their technology investments, but also structural damage to corporate culture and employee trust.
- Monitor not only the effects on productivity but also employee sentiment, collaborative behaviour, and mutual helpfulness. These ‘soft’ factors predict ultimate AI adoption better than actual technical performance.
- Help employees evolve their professional identity from “replaceable by AI” to “enhanced by AI.” This requires targeted coaching and career guidance.
- Identify and preserve those domains where human capabilities remain superior: complex problem-solving, ethical judgements, creative synthesis, emotional intelligence…
- Create work processes that make optimal use of both human and artificial intelligence, whereby both partners can contribute their strengths.
- Invest heavily in AI literacy so that employees become well-informed partners rather than passive users. Good understanding reduces anxiety and increases adoption.
- Develop clear guidelines for responsible AI implementation that place workers’ rights and human dignity at the centre.
The future of work
“The question is not whether AI will transform work — that is already happening,” concluded Puntoni. “The question is whether this transformation will be characterised by human flourishing or mass displacement and alienation.” A successful AI economy is not about technological supremacy, but about organisations that integrate technology and psychology simultaneously and optimally.
Carine Van Oosteren from TNO found Puntoni’s message very important: “Keep your eye on people in your AI transition. It is so easy to focus on the potential benefits in the form of productivity gains or a contribution to labour market shortages. But if you have insufficient attention for people, you will probably not achieve those benefits. Could this be the explanation for the very limited results from the deployment of genAI thus far?”
At TNO, she is researching the impact of social, economic, and technological developments on work, with specific attention to the future of work and the role of AI. She therefore confirmed the relevance of Puntoni’s message for business leaders: “Invest as much in the human side of AI as in the technical side, or you risk your AI strategy imploding through internal resistance.”
Follow-up discussions
The following day, various stakeholders discussed concrete policy pathways and European regulatory frameworks during a policy-focused gathering in Brussels. The afternoon began with a panel featuring representatives from the European Commission, LinkedIn, the Council of Europe, and Eurofound.
Carine Van Oosteren herself was subsequently responsible for one of the roundtable discussions typical of Reshaping Work: “I had deliberately chosen a provocative proposition: ‘the use of AI makes us lazy, dependent, and ultimately even less intelligent’. My table companions agreed with this. Many therefore had a love-hate relationship with the use of generative AI: it is very handy, but do I actually want to use it? When yes and when not? Some shared that their organisations had launched their own GPT models and that they are not allowed to use other AI models…” Through algorithmic management, amongst other things, AI is thus a major game changer for everything related to work. And this will only increase. We have not yet seen the end.
One conclusion stands firm: There is a need for more sustainable spaces where different societal actors can engage in open dialogue, not only about regulation, but also about daily practices, risks, and opportunities of AI at work. In that sense, this latest initiative by Reshaping Work was certainly successful!