THREAT-INFORMED ASSESSMENT
AI Threat Profiles by Industry
Generic AI security advice is a vendor problem, not a practitioner solution. Below: ATLAS-grounded adversary profiles tailored to specific verticals, each mapped to the AISS sub-controls that mitigate the threats — auditable end-to-end.
ROADMAP
Coming soon
Profiles in active research. If you operate in one of these verticals and want to contribute insight, open a Discussion on the AISS repo or reach out directly.
Financial Services
Fraud-detection evasion, trading-model extraction, supply-chain compromise of underwriting models, agentic back-office credential harvesting. 10 ATLAS techniques for FinServ AI.
PlannedLegal
Hallucinated case citations under bar discipline, privilege-bypass via prompt injection in document review, RAG poisoning of case-law corpora, agentic billing/filing-tool credential harvesting. 10 ATLAS techniques mapped to legal AI surface.
PlannedRetail & E-commerce
Recommendation-engine manipulation, pricing-model evasion, customer-data exfiltration via agentic checkout flows, supply-chain compromise of product-tagging models. Coming soon.
PlannedSee how your AI scores against these threats
An AISS assessment surfaces the gaps in audit-grade detail with transparent scoring math published at github.com/Ayliea/aiss.