Novo Nordisk has confirmed a major breach in which attackers stole large amounts of internal data, including AI models, training code, datasets, logs, and container images tied to its drug-discovery work. The incident also exposed pseudonymized clinical trial patient data and has triggered an investigation with Danish regulators and outside cybersecurity experts. #NovoNordisk #Dragonfly #vxunderground
Keypoints
- Attackers exfiltrated fully trained AI models and proprietary research data from Novo Nordisk.
- The stolen project, codenamed Dragonfly, included multimodal inference-ready model assets.
- Training codebases, logs from 113 runs, and internal infrastructure details were also taken.
- Pseudonymized clinical trial patient data was among the information confirmed stolen.
- The attackers are now extorting Novo Nordisk with a ransom demand.
Read More: https://securityonline.info/novo-nordisk-data-breach-ai-models/