AI‑detection startup Pangram announced a $9 million Series A round led by a consortium of venture firms, earmarked for expanding its proprietary detection engine and deepening its partnership with publishing platform Substack. Pangram’s technology scans text and multimedia for synthetic signatures, flagging content generated by large language models with a confidence score that can be embedded directly into a writer’s feed. Substack’s integration, rolled out this week, allows authors to display a “verified human” badge next to their bylines, giving readers a transparent cue about the origin of the material. The move addresses growing concerns over misinformation and the erosion of trust in digital publishing, positioning Substack as a pioneer in content authenticity.
The funding round underscores investor confidence in the emerging “trust layer” market, where startups provide verification tools to combat AI‑generated disinformation. Analysts estimate the sector could attract upwards of $2 billion in capital by 2028 as regulators tighten disclosure requirements. Pangram’s roadmap includes real‑time API services for social media platforms, expanded multilingual detection, and a marketplace for third‑party verification plugins. While the company boasts a 95 percent detection rate in internal benchmarks, critics caution that adversarial model fine‑tuning could erode efficacy over time, prompting a need for continuous model updates. Nonetheless, the Substack partnership serves as a proof point for commercial viability, and the infusion of capital will likely accelerate product rollout, positioning Pangram as a key player in the nascent AI‑integrity ecosystem.
This story was written by a language model from reporting by TechCrunch. Check the original before quoting it.