eva // weekly legal tech digest
← Topics·2 sources

AI Security & Resilience

Agentic risk, prompt injection, supply-chain attacks, and the encryption transition (including post-quantum) as it lands on legal infrastructure.

Current understanding

AI Security & Agentic Risk with Rok Popov Ledinski is the corpus's main security piece: agentic systems introduce attack surfaces that traditional InfoSec tooling doesn't address well — prompt injection through documents, tool-call abuse, agent-to-agent traffic. The framing: this is a discipline of its own, not a sub-bullet under existing security programs. The Quantum Paradox & the Race to Get Encryption-Ready runs in parallel: Google's claim that it can break RSA encryption by 2029 means "harvest now, decrypt later" attackers are already collecting encrypted legal data. The corpus treats this as a calendar problem, not a hypothetical — by the time post-quantum encryption is ubiquitous, today's TLS-protected exchanges are already cataloged. These two pieces don't share an author or audience but they share a posture: the security stance that worked five years ago does not work now, and law firms in particular have weak defenses against both threats.

Tensions

Mino relevance

Compliance and data residency are Mino-native moats: Dutch infrastructure, EU jurisdiction, traceable training data. For agentic security specifically, Mino's narrow-scope agent design *is* a defensive posture — small action surface, fewer tool calls, less prompt-injection blast radius. Worth saying loudly when selling into firms that have a CISO. Long-term watch item: post-quantum readiness will become a procurement check before most firms are ready, and Mino can be early.

Sources

2

Related