ETSI takes on AI cybersecurity
ETSI takes on AI cybersecurity
https://www.mobileworldlive.com/ai-cloud/etsi-takes-on-ai-cybersecurity/
Publish Date: 2026-01-15 11:27:00
Source Domain: www.mobileworldlive.com
Using an unordered list, summarize the following article with between 4 and 8 key points.
ETSI released its first globally applicable European Standard (EN) for AI cybersecurity, a move addressing the unique protection challenges posed by the technology by establishing baseline specifications.
The standards organisation stated EN 304 223 covers AI models and systems. It builds on principles established in a prior specification through a set of 13 doctrines and requirements covering secure design, development, deployment, maintenance and end of life phases.
Scott Cadzow, chair of ETSI’s technical committee for securing AI, said EN 304 223 is “an important step forward in establishing a common, rigorous foundation for securing AI systems”.
He noted a hike in the use of AI in “critical services and infrastructure” necessitates “clear, practical guidance” on security.
ETSI explained EN 304 223 received approval from various National Standards Organisations, giving it a global scope.
Subscribe to our newsletter
Get breaking news, exclusive insight, and expert analysis – before anyone else.
It identified data poisoning, model obfuscation, indirect prompt injection, and vulnerabilities resulting from complex data management and operational practices as key AI cybersecurity risks.
ETSI explained the five phases covered are in sync with globally recognised “AI lifecycle models, ensuring consistency and interoperability with existing standards and guidance”.
It expects EN 304 223 to “be instrumental for stakeholders throughout the AI supply chain”.
The specification’s scope covers deep neural networks and generative AI: ETSI plans to bolster the standard with a domain-specific arrangement for the latter element with a focus on deepfakes, mis or disinformation, confidentiality risks, copyright and IPR matters.
