Introduction. Under conditions of digital hyperpublicity, account-recovery methods based on a user’s “personal knowledge” become increasingly vulnerable. The mechanism of security questions, which relies on recalling autobiographical facts (a biographical code), faces a dual risk: answers can be compromised through open sources, and recall reliability can decline due to the reconstructive nature of autobiographical memory, interference, and the loss of contextual cues.

Materials and methods. An online quasi-experiment was conducted (N = 150; ages 20–45), including an assessment of digital engagement using the Digital Footprint Index and a cognitive test measuring recall of answers to 10 security questions of two types (standard and personalized). Accuracy, response time, and subjective confidence were recorded; for part of the sample, self-reports were selectively verified and a post-survey examined reasons for errors.

Results and discussion. Personalized questions were recalled worse than standard ones: lower accuracy, a higher proportion of “I don’t remember” responses, longer response times, and lower confidence; the differences were statistically significant (p < 0.001). A negative association was found between digital footprint volume and recall success (accuracy), and a positive association with response time; the effect was most pronounced for personalized questions. The post-survey indicated a significant contribution of interference and competing “versions” of memories.

Conclusion. The findings confirm systematic limitations of traditional security questions under digital openness and point to the need to shift toward more robust authentication mechanisms that incorporate contextual and behavioral user features.