Repetitive responsible gambling slogans and imagery may not be better than AI-generated content

Responsible Gambling (RG) messaging is a common element in online gambling harm reduction, but current practices may be suboptimal. Artificial Intelligence (AI) technologies enable rapid development of ‘always new’ RG content, but it is unclear whether such materials could be effective. Online gamblers from Australia, Canada, the UK, and the USA (N = 657, Mage = 38.93, cisgender male = 59.97%) played a simulated electronic gambling machine with (or without) in-play RG popup messages. Popups were comprised of either real (but unfamiliar) RG content or fully AI-generated content. Some participants saw ten different popups, while others saw the same popup repeated ten times. Popup exposure was associated with lower gambling cravings (t(652) = −3.10, p = .01, η2 = .02) and intentions (t(652) = −2.18, p = .03, η2 = .02). Participants exposed to changing popups reported lower gambling intentions (F(1,498.54) = 5.24, p = .02, η2p = .01) and higher message credibility (F(1,515) = 10.55, p = .007, η2p = .02). AI-generated messages were more recognizable (F(1,279.36) = 29.54, p < .001, η2p = .07) and less discouraging of gambling in general (F(1,515) = 12.40, p = .002, η2p = .02), but otherwise did not significantly differ from real RG content. These initial results suggest non-repetition may be important for RG content efficacy, while origin (human- or AI-generated) may be less important. Pending replication, operators and regulators could consider scaling ongoing efforts to deploy more varied RG content using AI-generated materials.