Just 4 percent of businesses achieved a return on their AI investments, yet rather than admit AI isn't living up to early expectations, a newly published study is blaming the users for not doing enough.
Based on research conducted by Economist Impact, part of the Economist media organization, the report says that business leaders give lip service to the strategic value of AI while neglecting to invest in the necessary skills, and few firms have a governance framework in place.
The research surveyed 639 senior decision‐makers in London, New York, Singapore, Sydney, and Tokyo to discover how organizations are preparing for an AI-powered future. It focused on three key areas – investment, developing talent, and culture & governance.
However, it found most companies are lacking in each area. Only 38 percent have a dedicated budget for investing in AI development and are struggling to embed the tech into core business processes.
Almost every executive surveyed claimed they are developing AI skills in their staff, but the report says most organizations rely on informal or ad hoc approaches such as mentorship or self‐directed online courses. Structured internal training or partnerships with external training bodies are reported by just 16 percent and 21 percent of the firms respectively.
As for responsible AI, many leaders claim to take this seriously, yet only 8 percent have implemented a comprehensive governance framework. Without clear standards or oversight, employees are left to manage AI risks alone, which could prove damaging in future, the report says.
The point the authors are trying to relay is that if AI is capable of doing what its proponents noisily claim, then it will introduce a structural shift in how companies operate and organize work. Customers likely to be successful will combine the best tools with a workforce ready and prepared to use them safely, it states.
This means matching technological investment with human investment and developing the necessary skills requires more than one-off training initiatives. All of this is bad news for managers who believe they can replace staff with some off-the-shelf AI - a trend highlighted last year.
Economist Impact also points toward a disconnect between top brass and middle managers when it comes to driving an AI strategy. Senior execs claim they champion AI talent, but nearly half of middle managers have minimal responsibility for AI skill development of their teams and 8 percent say they have none.
A third of top execs claim that resistance to change from employees and middle managers is a major barrier to aligning talent strategy with AI goals.
The report also notes that as AI automates routine work, human judgment becomes critical. This is because there isn't really anything intelligent about AI, and systems don't have any concept of what they are meant to achieve, so AI agents get office tasks wrong much of the time.
The bad news is that only a third of respondents claim their human workers excel at critical thinking and creativity, which limits innovation and could inhibit human oversight of AI-driven decisions.
What the research doesn't address is the thorny problem of how staff are supposed to maintain a high enough skill level to spot when AI has made a mistake. If developers aren't coding on a day-to-day basis, for example, will they still be able to identify flaws in AI-generated code?
This issue of skill erosion was raised last year at the Gartner Symposium conference in Australia, and Microsoft recently stated its concerns that AI may hollow out the profession's future skills base.
Economist Impact warns that small firms risk being left behind by the AI transition, if it turns out to be a game changer and not just another fad. Execs of these companies are more likely to cite budget constraints as a barrier to training, and nearly two-thirds of those surveyed say they lack the funds to hire specialists. The report claims AI will deepen the rift in capabilities between large and small firms without some form of targeted government support.
However AI may not turn out to be the panacea it is being promoted as. A larger recent survey of nearly 6,000 execs in the US, UK, Germany, and Australia found that more than 80 percent saw no discernible impact from using the technology over the past three years, on either employment or productivity at their firms. ®
Source: The register