Artificial Inequality: AI is exacerbating career, income, and gender divides, research from The Adaptavist Group reveals

Based on a survey of 4,000 knowledge workers across the UK, US, Germany, and Canada, the study reveals that higher earners have disproportionate access to the latest AI tools and training, allowing them to reap AI's promised rewards. In contrast, lower earners and women are being shut out from AI opportunities, which impacts their skill development, job satisfaction, and time savings, both personally and professionally.

As AI adoption continues to surge, this imbalance risks creating a lasting negative impact on income and career mobility if left unaddressed.

The findings show that respondents with household incomes of over £100,000 were more than twice as likely (27% versus 11%) to have received over 20 hours of AI training in the last year, compared with those on household incomes of £30,000 or less. As a result, 58% of those bringing in £100k+ strongly believe they've received sufficient guidance on AI, compared to 25% of those on less than £30k. More than three-quarters (78%) of those with six-figure incomes also said they were provided with access to new AI tools regularly, in stark contrast to less than half (49%) of respondents with incomes less than £30k.

This higher level of access to AI training is paying dividends for high earners:

While organisations stand to gain significantly by helping increase efficiencies and skills among the highest earners - traditionally the most senior resources - The Adaptavist Group's findings nonetheless raise urgent concerns about worsening workplace inequalities. The study provides compelling evidence to support Bhaskar Chakravorti's thesis of "artificial inequality", which highlights how AI makes societies' existing inequalities worse by "concentrating socioeconomic opportunities and outcomes within narrow societal segments while depriving others."

AI training divides are also emerging between large enterprises and small businesses. In fact, 24% of small businesses (1-50 staff members) have had no training at all in the last 12 months, and 56% have had less than three hours of training. Comparatively, just 12% of large organisations (more than 5,000 staff members) had received no training, whereas the majority (64%) had received more than three hours.

The training divide is particularly worrying given Digital Etiquette 2025 revealed a clear link between AI training and enablement and better outcomes. The study indicates that AI training is becoming a key differentiating factor for career mobility and success in the workplace, which highlights the need to support workers at all levels. Employees who received more than 20 hours of AI training were over four times more likely to view it as indispensable (29%) compared to those with an hour’s training or less (7%).

More training also translates to greater efficiency: 47% of those with 20 hours or more of training annually are saving at least 11 hours a week by using AI, equivalent to 1.4 working days for the average full-time employee. Of this figure, 14% said they are saving more than 30 hours, close to the equivalent of four full working days. Comparing this with those who had received less than an hour of training, just 7% said they were saving more than 11 hours a week, while a majority (58%) were saving less than 3 hours.

Those with more training not only saw more value but felt more comfortable proving it. Three-quarters (77%) of those with over 20 hours of training felt comfortable proving the ROI of their AI tools, compared to 16% of those with an hour or less of training. There is also a direct correlation between training and job satisfaction. Some 53% of those with more than 20 hours of training said AI had significantly increased their job satisfaction, versus a meagre 8% of those who'd received an hour's training or less.

Figures suggest that company culture may be hindering AI skills and outcomes, as one third (35%) of workers said they wanted more training but were afraid to ask for it.

The study also uncovered that women received less training than their male counterparts, despite the clear correlation between training and AI value:

This is borne out across the organisational hierarchy. Women are given less access to training than their male counterparts across organisations, with very few exceptions.

Only 58% of women in director roles received structured training sessions on AI from external providers compared with 73% of men in the same position. At intern level, the figures show that men are more than twice as likely to have received external training (47% versus 23%).

Amongst administrative staff, 35% of women received structured external training compared with 52% of men. At a time when AI is increasingly being used to automate administrative tasks, fueling job cuts, this poses an urgent question about whether women will be more exposed to redundancy risk than men.

Asked whether they believed they had been given 'sufficient guidance on how to use AI at work', there is a further gap between men (87%) and women (77%) at C-level. This trend holds amongst administrative staff, with 46% of women citing sufficient guidance compared with 57% of men. This reveals gaps in skills perception and AI awareness, as 72% of men believe their skills are developing because of AI, compared to 63% of women.

"The findings clearly show how access to training and tools enables professionals to unlock value with AI, while simultaneously increasing job satisfaction.

"With so much to be gained from properly implemented AI, it is critical that opportunity is not concentrated in the hands of the few. Organisations should take steps to ensure equal access to tools and training across the workforce. Without this, we risk deepening the very inequalities technology should help to solve.

"For instance, the advantages of AI also extend beyond the workplace. Our study uncovered that higher earners are significantly more likely to use AI both professionally and personally. While we didn't explore personal usage patterns, AI's potential to reduce unpaid labour burdens and support side hustles suggests that unequal access to AI training could deepen existing divides, particularly problematic in an economy that increasingly values AI proficiency."

"Popular AI tools are built in the vision of Silicon Valley and best serve its demographics, leaving others behind. This critical research exposes how higher earners with more AI training extract more value, and explicitly shows the gender disparity in similar roles, with women facing reduced access and opportunities. As AI adoption accelerates, this creates a divide between AI 'haves' and 'have-nots' which will deepen inequality at work.

"This isn't just about fairness, it's about organisational survival. Companies that prioritise inclusive AI principles will outperform those that don't. The choice is simple: give everyone AI access for exponential growth, or concentrate opportunity among the few and limit your potential."