AI-SPM vs OpenAI Privacy Filter

AI-SPM (AI Security Posture Management) is an open-source infrastructure layer for securing LLM pipelines running in production. It targets three attack surfaces that traditional application security doesn't cover: prompt injection (including obfuscated and multi-step variants), tool abuse via unvalidated structured outputs, and data exfiltration through PII leakage in model responses. The architecture layers a gateway intercept layer over incoming prompts, runs context inspection before the LLM sees any input, enforces policies via Open Policy Agent (OPA) for declarative, auditable rules, then pipes all events through Apache Kafka and Apache Flink for real-time streaming analysis. This means security posture can be monitored and enforced at scale without blocking the inference path. The project is genuinely fresh — posted as a Show HN today. Early community feedback pointed to capability-based token models (similar to OS kernel permission rings) as a complementary approach to content-scanning, which the author acknowledged as a meaningful gap. The timing is right: as companies push AI agents from demos to production, the security tooling layer is largely underdeveloped. AI-SPM is one of the first OSS projects to tackle it at the infrastructure layer rather than with prompt-level guardrails alone.

OpenAI released Privacy Filter on April 22, 2026 — a 1.5B-parameter open-weight model for detecting and redacting personally identifiable information from text before it ever reaches a cloud API. The model runs fully locally, handles 128,000 tokens in a single pass, and achieves a 96% F1 score across eight PII categories: names, addresses, emails, phone numbers, URLs, dates, account numbers, and secrets. Unlike traditional regex-based PII scrubbers that choke on unstructured text and context-dependent references, Privacy Filter uses a fine-tuned language model to understand semantic context — it catches "call me at the usual number" type references that pattern matchers miss entirely. The model ships with only 50M active parameters at inference time via sparse activation, keeping latency low enough for preprocessing pipelines. Available on Hugging Face and GitHub under Apache 2.0, Privacy Filter solves a real bottleneck: enterprises and regulated industries have been unable to safely pipe sensitive documents through LLMs at scale. OpenAI explicitly warns it should be treated as a "redaction aid, not a safety guarantee," which is unusually honest for a model card — and a sensible framing for high-stakes medical or legal workflows.