On January 14, 2025, the Cybersecurity and Infrastructure Security Agency (CISA) released the AI Cybersecurity Collaboration Playbook (the “Playbook”) to provide guidance to organizations within the AI community (including AI providers, developers, and adopters) to voluntarily share AI-related cybersecurity information with CISA and its partners through the Joint Cyber Defense Collaborative (JCDC). To combat AI-related cybersecurity threats and enhance the cybersecurity resilience of AI systems, the Playbook recommends organizations to incorporate its recommendations into their existing practices.
Goals of the Playbook
The Playbook aims to achieve the following goals:
- Increase collaboration between federal agencies, international partners, private industry, and other stakeholders to improve the resilience of AI systems and facilitate awareness of AI cybersecurity risks;
- Guide JCDC partners to voluntarily share information involving AI-related cybersecurity vulnerabilities and incidents;
- Outline safeguards for information sharing mechanisms and protections; and
- Explain CISA’s actions upon receiving shared information to strengthen collective defense.
Information Sharing Protections under CISA 2015
While information sharing is entirely voluntarily, the Playbook outlines the various legal protections afforded under the Cybersecurity Information Sharing Act of 2015 (CISA 2015) through participation as a JCDC partner, including the protection of proprietary information, exemption from disclosure under Freedom of Information Act (FOIA), prohibition on the use of such information in regulatory enforcement, and non-waiver of privilege.
AI-related information may also be covered under CISA 2015 if the information qualifies as a cyber threat indicator or defensive measure, which is further detailed in a DHS-DOJ guidance.
Recommendations for Information Sharing Framework
The Playbook highlights the importance of transparent and frequent information sharing to address AI-related risks and threats. Key recommendations include:
- Proactive Sharing. JCDC partners are encouraged to share information on malicious activities, emerging trends, and assessments to help quickly identify and mitigate critical threats.
- Incident or Vulnerability Reporting. JCDC partners are urged to actively share information related to an AI cybersecurity incident or vulnerability, which may be shared through CISA’s webform to report a cyber incident or product or service vulnerability. If using a web form, CISA also recommends notifying a JCDC representative through email.
- Information Analysis. Information shared by JCDC partners will be aggregated, validated, analyzed, anonymized, and enriched by CISA to determine defensive actions and shared with various partners, including government, private sector, and international stakeholders.
Defensive Actions by CISA
Upon receiving and assessing the information shared by JCDC partners, CISA may take the following actions in collaboration with multiple stakeholders, depending on the nature of the identified threat or vulnerability:
- Identify and notify victim organizations;
- Offer proactive engagements and services;
- Provide information on adversary tactics and infrastructure;
- Share information for detection and prevention purposes; and
- Distribute relevant threat intelligence products.
Call to Action
The Playbook marks a significant step towards creating a collaborative defense against evolving AI-related cybersecurity threats. As organizations continue to use and rely on AI technologies, the Playbook highlights the importance of shared cooperation and vigilance to mitigate AI risks.
Given the rise in AI threats, the Playbook encourages organizations to consider identifying priority issues for the AI community, promoting post-mortem knowledge sharing and analyses, flagging opportunity for technical exchanges regarding emerging AI threats, adversaries, and vulnerabilities, and becoming a JDCD partner for collaborative cybersecurity planning and defense.
