Adoption of workplace AI is inconsistent and riddled with contradictions

New research claims to reveal the significant divides in how organisations are adopting artificial intelligence, with some reporting positive outcomes and others citing fear, risk and job losses. The study, carried out by The Adaptavist Group as part of its latest Digital Etiquette: Unlocking the AI Gates report, surveyed 900 professionals responsible for introducing and onboarding workplace AI in the UK, US, Canada and Germany. It found a sharp split between the 42 percent of respondents who believe their company’s AI claims are over-inflated and the 36 percent who do not.

Leaders who view their company’s AI claims as exaggerated expressed far greater concerns about the technology. Sixty five percent of this group said they believe their organisation’s stance on AI puts customers at risk of financial, psychological or physical harm, compared to just 9 percent of those who consider their company’s claims realistic. Despite these concerns, spending on AI remains high across both groups. Thirty four percent of sceptical leaders reported investments of between £1 million and £10 million over the past year, while 8 percent said their company had invested more than £10 million. This compares to 31 percent and 6 percent respectively among the more optimistic group, indicating that attitudes are not linked to investment levels.

Pressure to adopt AI was also more pronounced among sceptical leaders. Eighty four percent said they encourage their teams to use the technology because they feel they should, rather than because it delivers specific value, compared with 58 percent of those in organisations with more measured expectations. Respondents also reported higher levels of workplace fear and restrictions. Two thirds of sceptical leaders worry about AI adoption putting jobs at risk, while 80 percent said their company is using AI innovation to reduce headcount. In contrast, only 10 percent and 39 percent of the more positive group expressed the same concerns.

It’s the culture

Cultural impacts are also evident. Forty two percent of sceptical leaders said they hide their AI usage at work, compared with 6 percent of those in companies taking a more realistic view. Nearly half of respondents in the sceptical group said they feared being wrongly accused of using AI, while 45 percent believed colleagues see AI users as less competent. The figures for those with more positive views were 15 percent and 5 percent. Sceptical organisations were also more likely to tie AI usage to performance targets and discourage experimentation, with 50 percent linking AI use to KPIs compared to 28 percent of realists, and 66 percent discouraging experimentation compared to 16 percent.

A lack of training appears to reinforce these problems. Fifty nine percent of leaders in organisations accused of overinflating AI claims said they had no access to formal training, compared to 16 percent in more measured environments. Previous research from The Adaptavist Group predictably highlighted the importance of training in improving outcomes.

Leaders in organisations where AI was introduced with greater clarity and training reported stronger benefits. Fifty eight percent said AI had improved work quality, 61 percent cited time savings and 48 percent noted increased output. Concerns about plagiarism, bias, hallucinations and inaccuracies were also significantly lower among this group. Only 37 percent of realists reported worries about plagiarism compared with 74 percent of sceptical leaders, and just 21 percent expressed concerns about hallucinations versus 72 percent of sceptics.

Sceptical leaders also reported greater frustration with AI outputs. Seventy four percent said they regularly edited or regenerated responses, compared with 25 percent of the more positive group. Rapid introduction of new tools has further added to the divide. Seventy four percent of sceptical leaders said too many AI tools were being introduced too quickly, while only 24 percent of realists agreed.

The outlook for future adoption also differed. Sixty seven percent of sceptical leaders expect their AI usage to decrease in the next year, compared to just 9 percent of those in companies with more realistic expectations.