I am increasingly uncomfortable with how quickly the conversation about AI becomes a judgment about people.

They are not learning fast enough. They are resistant. They need to become more adaptable.

Sometimes there is a skill that needs to be developed. Sometimes a person needs to try something unfamiliar. We all do.

But before we decide that someone is unwilling to change, I want to know what support was actually available.

Were they given time to learn? A tool they were allowed to use? A practical connection between the training and the work they do every day?

Or were they simply told that the future had arrived and keeping up was now their responsibility?

The expectations are rising. The support is not keeping pace.

PwC’s September 2026 Global Workforce Hopes and Fears Survey offers a reason to examine this more closely.

Among nearly 50,000 workers across 48 countries and regions, 64% reported using AI at work during the previous year. That was up ten percentage points. Yet the share who said they had access to the learning and development resources they needed fell from 59% to 51%.

That learning measure covers development generally, not just AI training. Even so, the direction deserves attention.

More people are using AI. Fewer say they have the support they need to develop.

Source: PwC, Global Workforce Hopes and Fears Survey 2026 ↗

You cannot keep raising expectations while leaving people to figure out the preparation on their own, then treat the uneven results as a fair measure of talent.

My experience in adult learning and corporate software training makes me particularly attentive to this. Knowing that a tool exists is one thing. Understanding how to use it well in the middle of an actual workday is another.

A person who has spent years recognizing problems, managing exceptions and helping others through change brings something valuable to the learning process. Their experience should be part of the design.

Who gets the opportunity to practice?

Imagine two employees.

One receives a paid account, time to experiment and a manager who asks what they are discovering. The other receives a link to a webinar and the same full workload they had yesterday.

A month later, we compare their progress.

What exactly are we measuring?

This is an illustrative example, but it is the question I would bring into an organizational conversation. Access includes the conditions that make learning possible: time, relevant work, clear guidance and someone who can help when the first attempt goes badly.

Gender, race and location belong in that conversation. They affect what we need to examine. They should never become shortcuts for deciding what someone is capable of doing.

Gender: do not confuse a difference in use with a lack of ability

The familiar story that women are simply behind on AI is too broad.

Pew Research Center’s June 2026 findings showed similar overall chatbot use among U.S. men and women: 50% and 47%. There were differences in frequency and workplace use. Among employed adults, 40% of men and 35% of women reported using chatbots for work tasks.

Those figures tell us about use. They do not tell us why someone uses a tool less, what training they received or how capable they are.

Source: Pew Research Center, The gender gap in AI ↗

The work people are doing also matters. The International Labour Organization’s March 2026 analysis found greater exposure to generative AI among women than men in 88% of the countries analyzed, connecting that pattern to the occupations in which women are concentrated.

Exposure means tasks could be affected. It does not mean those workers have already lost their jobs.

Source: International Labour Organization, Women and generative AI at work ↗

If administrative and support work is being redesigned, I want to know whether the people who understand that work are being prepared for what comes next.

Are they helping shape the new process? Are they receiving useful training? Can they see a path into a different responsibility?

Telling someone to be more confident does very little if the opportunity to advance remains unclear.

Race: ask better questions before telling the usual story

As a Black woman, I pay attention to how quickly a conversation about a gap can become a conversation about what a group supposedly lacks.

The evidence here deserves more care.

In Pew’s 2026 U.S. survey, 36% of employed Black adults and 36% of employed White adults reported using chatbots for work. The figure was 37% among Hispanic adults and 60% among Asian adults; the Asian estimates represent English-speaking adults only.

That does not support a blanket assumption that Black workers are less engaged with these tools than White workers.

Source: Pew Research Center, Racial and ethnic differences in AI use ↗

But equal reported use does not answer every question about opportunity.

Who gets the advanced training? Who is invited into the project that leadership will notice? Who receives credit when a new approach works?

These are questions organizations need to investigate. The survey does not establish those answers.

And separate race and gender averages cannot tell us the specific experience of Black women. We need enough curiosity to look beyond an overall number and understand what people are actually encountering.

A capable professional should not have to overcome a stereotype before their work can be evaluated fairly.

Location: being online does not put everyone in the same position

We sometimes talk about digital tools as though geography has stopped mattering.

It still matters.

The OECD’s 2024 regional analysis estimated that 32% of urban workers were in occupations exposed to generative AI, compared with 21% of rural workers. It warned that the technology could widen existing regional productivity and digital divides.

Those are exposure estimates, not current adoption rates or job-loss figures. They help show why the same technology can arrive in very different local circumstances.

Source: OECD, The geography of generative AI ↗

For an organization, I would examine whether a regional office has the same useful support as headquarters. Whether training works across languages and time zones. Whether people in smaller teams have anyone available to answer their questions.

A session being available online does not necessarily make it accessible during someone’s working day.

For the individual considering global living, this matters too. A move can create new possibilities, but your learning opportunities, professional relationships and access to useful support still deserve a place in the plan.

Experience needs a place in the transition

I do not want experienced professionals to spend this period apologizing for having learned their work before the latest technology arrived.

You still need to learn. Your experience also gives you something to contribute.

You may recognize when an answer is technically polished but practically wrong. You may know which exception will derail a process. You may understand why a change that looks efficient creates more work for everyone downstream.

The task is to connect that judgment with the tools and make the resulting value visible.

Organizations can help by giving people real problems to work on, paid time to practice and a fair opportunity to demonstrate what improves. Look beyond the employees who already appear most comfortable with technology.

Professionals can begin with one recurring problem they understand well. Use an approved tool and non-sensitive material. Evaluate the result. Keep a clear example of where your judgment made the difference.

The opportunity to build that example should not depend on being in the right office, having the right manager or already being seen as someone who belongs in the technology conversation.

Organizations need to examine who receives useful tools, paid practice time and a chance to apply what they learn. Looking at gender, race and location helps reveal where the conditions differ. It does not give us permission to make assumptions about ability.