The paper looked over today from Peikai Li (et al.) from the Leeds University Business School introduces the concept of SUGAR or Stealth Use [At work] of Generative AI Resources.
For a number of reasons employees (and not a few academics it’s also noted!) are using generative AI stealthily.
It may be they fear jobs being replaced if they acknowledge this help. It may be they perceive a competitive advantage over peers and are reluctant to share methodology. Or, it may simply be a hack that saves time but they don’t want to acknowledge they’re thus free for more demanding tasks.
Whatever, SUGAR is real and the consequences are not good [N.B. the study was conducted using subjects in both China and the UK which helps to rule out cultural bias.].
To paraphrase the work’s main conclusion: SUGAR produces a negative relationship with task performance as subjects suffer reduced reflective learning and heightened feelings of ‘Imposter Syndrome’. In English? People don’t like how they feel about themselves when they’ve completed a task if SUGAR has been used. Moreover, the effort of concealment adds an additional stress to working lives which leads to diminished overall performance and less useful employees.
The study found around 33% of respondents surreptitiously using gen. AI (but I’d guess the real number is higher). Employers are recommended therefore to clarify policies as a) they’ll have happier workers if they do and, b) the firm will be safer if it’s not leaking information out of uncleared back-doors to unauthorized third parties.
The researchers end by noting it’s early days for such a study but they feel there’s a strong enough link between SUGAR and negative outcomes to justify keeping an eye on how this phenomena develops.
To get into the weeds follow the link How Stealth Use of Generative AI Resources Impairs Task Performance.
Happy Sunday