Social engineering statistics

Social Engineering Examples, a Diopter AI research project·Recomputed 11 Aug 2026

What the corpus shows

• Diopter AI's analysis of 173 documented cases records phishing as the entry channel in 62 of them, more than any other technique.
• Across the corpus, Diopter AI finds the most-targeted sector is Financial Services & Insurance, with 46 cases.
• Among the 77 cases that state a loss in US dollars, Diopter AI records a median of $2.3M. Half of all disclosed losses fall below that figure.
• The largest single verified loss is $540M, at Axie Infinity / Ronin Bridge Heist: A Fake LinkedIn Job Offer That Cost ~$600M.
• Diopter AI verified every case against primary reporting, citing 1117 sources in total, an average of 6.5 per case.
• Of the 173 cases, 170 are confirmed by the victim, a regulator or a court, and 3 remain alleged.

Cases by attack channel

A case can use more than one channel, so these are counts of cases, not shares of the total.

Cases by sector

Top ten sectors. A case can affect more than one.

How disclosed losses are distributed

The 77 cases that state a figure in US dollars.

How these figures are counted

A case can involve more than one channel and more than one sector, so channel and sector figures are counts of cases rather than percentages, and they sum to more than 173.

We publish a median rather than a grand total. The library records amounts stolen from one organisation alongside remediation costs, regulatory settlements and multi-victim campaign figures, and those are different quantities that cannot meaningfully be added together. Where a source gives only a modelled or economy-wide estimate, or states that no figure was ever disclosed, the case is counted as having no disclosed loss.

Figures stated only in a currency other than US dollars are excluded rather than converted at a rate we cannot source. Per-case sourcing is set out in our methodology.