Session
When the Disassembler Lies: Can AI Reverse Engineers Recover?
AI-assisted reverse engineering depends on the evidence produced by its tools. When a disassembler returns a plausible but incorrect listing, an agent can build a confident explanation on the wrong foundation. This session examines how obfuscation affects the loader, disassembler and decompiler, and how those failures differ from prompt injection in strings or tool output.
Using controlled binary-analysis examples, we examine how an agent can recognise an unsupported assumption and seek independent evidence. We compare explicit tool errors with misleading output that arrives without a warning, then consider recovery through another analysis view or runtime observation. A successful tool call and a correct explanation of the program are separate outcomes.
Attendees learn how to assess evidence in an agent's analysis trace, choose useful cross-checks and set limits on repeated or inconclusive analysis. The session is intended for practitioners evaluating AI-assisted reverse engineering and malware triage.
Audience: Reverse engineers, security researchers and teams building AI-assisted analysis tools. Binary-analysis basics are helpful.
Format: 60 minutes including demonstrations and Q&A. A focused 30-minute version is available.
Languages: English and German.
Anatoli Kalysch
CISO at Lindenhofgruppe | Swiss CISO of the Year 2026 | AI security and cyber risk
Bern, Switzerland
Links
Please note that Sessionize is not responsible for the accuracy or validity of the data provided by speakers. If you suspect this profile to be fake or spam, please let us know.
Jump to top