How Ageism Is Undermining AI Implementation
From Harvard Business Review, Feb. 25, 2025
by Mona Mourshed and Anika Heavener
https://hbr.org/2025/02/how-ageism-is-undermining-ai-implementation
Summary. Companies often assume that older workers are less adept at adapting to new technologies. As leaders look to bring new AI tools onboard, many believe that younger workers will be better able to integrate these tools into their work. A recent survey suggests this view may be unfounded. When hiring managers—many of whom held this view—were asked to assess the job performance of the mid-career and older workers whom they already employ, 89% reported that these individuals performed as well if not better than younger peers. Mid-career and older workers offer unique strengths when it comes to working with AI. For companies that want to drive AI adoption, these workers are an essential asset that can’t be overlooked.
======================
Employers are currently making two big assumptions. First, that AI tools such as generative AI and AI agents are going to remake how we work. And second, that mid-career and older workers aren’t suited to adapt to these technologies. While we’ll have to wait and see how the first assumption shakes out, we know right now that the second is wrong. And unless it’s challenged, both workers and employers will lose out.
As the founding CEO of Generation, a global nonprofit that trains and places adults of all ages into new careers (Mona), and as the head of innovation and investments for The SCAN Foundation, which supports older adults (Anika), we’ve seen this kind of pre-existing technology bias against older workers before. In Generation’s 2021 survey, when employers were asked to rate job candidate strengths, they overwhelmingly identified younger candidates as more application-ready, possessing more relevant experience, and as better fits with company culture relative to those who were age 45+. Similarly, in the Generation-OECD follow-up survey in 2023, 47% of hiring managers said they would hire a candidate age 30–44 for an entry- or intermediate-level role while only 13% said the same about candidates age 55+.
What’s troubling about these earlier judgments is how disconnected from actual results they were. As it turns out, when we asked those same hiring managers to assess the job performance of the midcareer and older workers whom they already employ, 89% reported that these individuals performed as well if not better than younger peers.
Unfounded age bias is likely to prove equally unhelpful now as employers navigate rolling out AI in the workplace. And according to our new survey, while midcareer and older workers offer unique strengths, employers are even more likely to overlook them than before in this new AI environment.
Ageism Is Undermining AI Implementation
Generation recently conducted a survey to explore the experience of midcareer and older workers in light of the growing use of generative AI tools in the U.S. and in four European countries: France, Ireland, Spain, and the United Kingdom. (This survey was supported with grant support from The SCAN Foundation in the U.S. and Google.org in Europe.) Our sample included 2,610 employees age 45+ who work in entry- and mid-level roles, and 1,488 hiring managers across a range of industries and company sizes who evaluate candidates for these roles.
Across the board, employers prefer younger candidates. Just 32% of employers in the U.S., we found, would “likely” consider candidates over age 60 for roles that regularly use AI tools versus 90% who would likely consider candidates under age 35. In Europe, 33% of companies said they were likely to consider hiring candidates over age 55 for such roles, while 86% were likely to consider the younger under age 35 group. The middle age range fell somewhere in between – with a declining interest as they got older (in the U.S., 79% of employers were likely to consider candidates aged 35–44, and 53% for those aged 45–59. In Europe, 71% were likely to consider candidates aged 35–44, and 52% for those aged 45–54).
Given the enthusiasm for adopting the latest generation of generative AI tools, this persistent bias bodes poorly for older workers. Nearly half of U.S. hiring managers in our survey reported that their companies have started providing AI tools to employees, and an additional 24% plan to do so in the next 12 months. While Europe’s transition is slower, change is underway there as well, with 29% of employers saying they currently provide AI tools and 31% planning to in the next year. Being able to use these tools is clearly a priority for companies and hiring managers.
However, it’s also possible that business leaders don’t yet know what they need to succeed with AI. Many companies report struggling with their AI transition. According to the LinkedIn Executive Confidence Index, 40% of global executives say their organizations have “limited leadership alignment, tools, and ad hoc processes in place” for generative AI adoption, while 25% say their organizations have not yet started this process.
Workers are also at sea. A recent survey by Upwork showed that employees across all ages and roles remain unsure how to use AI. Nearly half (47%) of employees reported that they have no idea how to use AI to achieve the productivity gains their employers expect and 77% saying these tools have actually decreased their productivity and added to their workload. Similarly, a Gallup Q2 2024 survey found that nearly seven in 10 U.S. employees say they never use AI, while only one in 10 say they use it at least weekly.
But these broad numbers obscure a critical bright spot that our survey identified. One key group of older employees, it turns out, are having a very different encounter with AI. Among the 15% of midcareer and older workers in our survey who say they have already adopted AI tools in the workplace, most are self-taught “power users.” This group turns to AI tools multiple times a week, if not daily. More than half of older power users in the U.S. and two-thirds in Europe reported that AI is improving their work quality and pace. In both the U.S. and Europe, nearly half of these power users added that the technology allows them to do more advanced or higher-level work, and about a third said it helps them make better decisions.
This small but important group of users shows a way forward for organizations that are implementing AI. But making progress will require employers to recognize that only a true inter-generational effort will enable AI to realize its full potential in the workplace.
A Two-Step Solution for Employers
To solve their AI transition problems, companies should take a fresh look at why they’re struggling—and the resources they already have. There is widespread alignment that a “human in the loop” is necessary to ensure that AI agents—the use of which is expected to increase dramatically in 2025—operate in an accurate, secure, and ethical manner. These AI tools are created by tech players or customized by companies to work autonomously to accomplish complex tasks that would otherwise be given to a human to perform. To that end, workers with years of business judgement and pattern recognition—such as midcareer and older workers—can be well-placed to play this role. But enabling this opportunity requires employers to be much more intentional about how to marry AI with worker experience.
Step One: Lean Harder on Experienced Power Users.
A starting point could be to double-down on leveraging the insights of the small group of older power users who have already taught themselves how to use AI tools. Spotlighting these employees and incentivizing them to share and mentor their peers on how to do the same can help employees of all ages. Moving a few into critical new roles where AI adoption is front and center can also help. Uncovering more relevant use cases that marry experience with AI tools, and then applying them to specific roles and tasks, should increase the odds that those much-touted productivity gains will be realized faster.
Step Two: Redouble Efforts to Educate and Engage the Rest—Young and Old Alike.
But making the most of experienced workers will require more from employers than merely better leveraging the AI proficiency of the already converted. As our survey also found, most midcareer and older workers are still on the fence about the value of AI. While 24% of non-users who are age 45 and over in the US and 36% in Europe expressed interest in learning to use AI tools, another quarter were neutral or unsure. Most troubling, nearly a third of the non-users in the U.S. and 17% in Europe declared they are not interested in AI at all.
To bring this larger group of experienced workers into the tent and tap into their institutional know-how, companies need to build out an important new capability. Rather than just rely on self-taught power users across the age spectrum, they also need to integrate AI skills into their formal training programs, while keeping the rapidly changing content relevant and continuously identifying relevant use cases. These programs will then need to better target the many midcareer and older workers who will need to reskill to move to a new job or upskill within a role. For their part, those workers will need to guard against falling back on experience alone as the best way to stay employable.
Microsoft shows how this two-step approach can work well. It has implemented company-wide upskilling and reskilling initiatives through platforms such as Microsoft Learn and LinkedIn Learning, designed to be accessible to employees of all ages. Notably, Microsoft’s internal programs often identify future “AI manager” roles that require both technical AI knowledge and strong leadership experience—areas where seasoned professionals’ judgment is especially critical.
In line with the growing need for responsible AI deployment, Microsoft has also placed experienced professionals in key roles guiding company policy, ensuring regulatory compliance, and managing potential risks. The company’s Industry Cloud solutions, which require deep domain expertise, often rely on the knowledge of midcareer and older workers to fine-tune AI applications for real-world scenarios. Furthermore, with the emergence of generative AI tools such as Microsoft 365 Copilot, Microsoft recognizes the need for seasoned managers to act as “managers of AI agents,” leveraging their institutional knowledge and critical thinking skills to oversee AI deployment, interpret outputs, and ensure alignment with business goals. Microsoft’s approach to AI training and integration demonstrates how companies can marry AI capabilities with the valuable experience of midcareer and older workers, creating a synergy that maximizes the potential of both human and AI across generations.
Actions like these can help find emerging answers that will give everyone—employees and employers—a greater ability both to better leverage this new wave of agents and to keep evolving our workplaces to harness AI’s transformative power.
