Axolile Lungu
Axolile Lungu is a qualified chartered accountant and the founder of synthetic cfo, which generates SAP and Oracle ledgers from accounting rules with fraud planted, labelled at the moment it is planted, and innocent look-alikes beside it, so that fraud detection can be trained and measured without a single real record. He audited multinationals in pharmaceuticals, technology, consumer goods and engineering at PwC in South Africa and EY in Ireland, worked in financial accounting and product costing at Gilead Sciences and AbbVie through an SAP S/4HANA implementation, and managed revenue accounting, SOX compliance and fraud resolution at Meta. Since 2025 he has advised AI companies in the United States, the United Kingdom and India on the forensic validation of financial data and the evaluation of finance models. He holds an MBA and is a dual citizen of South Africa and Ireland.
Writing
Six incidents, chosen by the company that had them
A few days ago I wrote that OpenAI had conceded it has no standard for when to disclose misalignment. That same day it published one. It builds two of the four things a disclosure regime needs, and the two it cannot build for itself are the two that do the work.
AI safety has discovered auditing, and auditing has notes
A researcher's resignation set off a week in which the frontier labs asked for embedded evaluators, tests a model cannot see coming, and a licensed profession to run them. Audit has done all of it for a century, and the hard lessons are the ones being skipped.
Label Provenance and Innocent Populations in Synthetic ERP Fraud Benchmarks
Labelled occupational fraud data is scarce, because identifying fraud inside real enterprise ledgers requires both known cases and expensive expert annotation.