EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

From Awareness to Resilience: Generative AI-Enabled Social Engineering and the Redesign of Security Education

Nicki James Shepherd

Journal of Artificial Intelligence General science (JAIGS) ISSN:3006-4023 · 01.12.2024 · 7 (01) · 444-459

DOI: 10.60087/jaigs.v7i01.515 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Artificial Intelligence General science (JAIGS) ISSN:3006-4023, online erschienen am 01.12.2024. Verfasst von Nicki James Shepherd. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Security awareness training rests on the assumption that users can be taught to recognise a malicious message by its flaws, and generative artificial intelligence (AI) removes most of those flaws. Before the end of 2022, researchers had already shown that large language models could write spear-phishing emails that outperformed those written by security professionals, that listeners could not reliably distinguish synthetic speech from real speech, that people rated AI-synthesised faces as more trustworthy than real ones, and that a cloned executive voice had been used to extract €220,000 from a UK firm. At the same time, large field studies found that embedded phishing training, the dominant form of training in industry, had little or no measurable effect on whether staff clicked. This paper argues that the goal of security education should move from awareness, meaning the ability to spot deception, to resilience, meaning the capacity of people and institutions to limit harm once deception has succeeded. Using a structured narrative synthesis and conceptual framework development, it proposes the Awareness-to-Resilience (A2R) framework. The framework has four layers: controls that remove decisions from users, verification routines that do not depend on detection, a reporting culture that treats users as sensors, and inclusive learning design. The last layer responds to evidence that cue-based training places particular demands on disabled learners, including screen reader users and people with dyslexia. The paper maps the framework against five attack types, sets out its implications for further and higher education, and identifies the empirical work needed to test it.

Zitieren

APA

Shepherd, N. J. (2024). From Awareness to Resilience: Generative AI-Enabled Social Engineering and the Redesign of Security Education. Journal of Artificial Intelligence General science (JAIGS) ISSN:3006-4023, 7(01), 444-459. https://doi.org/10.60087/jaigs.v7i01.515

Vancouver

Shepherd NJ. From Awareness to Resilience: Generative AI-Enabled Social Engineering and the Redesign of Security Education. Journal of Artificial Intelligence General science (JAIGS) ISSN:3006-4023. 2024;7(01):444-459. doi:10.60087/jaigs.v7i01.515

BibTeX
@article{Shepherd2024,
  author = {Shepherd, Nicki James},
  title = {From Awareness to Resilience: Generative AI-Enabled Social Engineering and the Redesign of Security Education},
  journal = {Journal of Artificial Intelligence General science (JAIGS) ISSN:3006-4023},
  year = {2024},
  volume = {7},
  number = {01},
  pages = {444--459},
  doi = {10.60087/jaigs.v7i01.515}
}
Kostenlos starten →