Sirion的动态

Forvis Mazars US provides details on new DOJ updates calling for corporations to have a mitigation plan to address potential AI risks. Review key takeaways and how they may affect your organization: https://lnkd.in/eGRpnu8a

Forvis Mazars US provides details on new DOJ updates calling for corporations to have a mitigation plan to address potential AI risks. Review key takeaways and how they may affect your organization: https://lnkd.in/gJ6hCy4d

Forvis Mazars US provides details on new DOJ updates calling for corporations to have a mitigation plan to address potential AI risks. Review key takeaways and how they may affect your organization: https://lnkd.in/eibx7Qyd

Forvis Mazars US provides details on new DOJ updates calling for corporations to have a mitigation plan to address potential AI risks. Review key takeaways and how they may affect your organization: https://lnkd.in/gR57Z9rU

Forvis Mazars US provides details on new DOJ updates calling for corporations to have a mitigation plan to address potential AI risks. Review key takeaways and how they may affect your organization: https://lnkd.in/evkj6vJ6

Forvis Mazars US provides details on new DOJ updates calling for corporations to have a mitigation plan to address potential AI risks. Review key takeaways and how they may affect your organization: https://lnkd.in/d_rpMAkm

Forvis Mazars US provides details on new DOJ updates calling for corporations to have a mitigation plan to address potential AI risks. Review key takeaways and how they may affect your organization: https://lnkd.in/ev-CWukD

The Office of Management and Budget (OMB) recently issued Memorandum M-24-18 to guide federal agencies on ethical and effective AI system acquisition. Focusing on three fundamental goals—managing AI risks, fostering collaboration, and promoting competition—these new requirements aim to improve accountability, privacy, and reduce market control by a few vendors. With this directive, agencies must create acquisition strategies emphasizing performance-based methods, transparency in licensing, and reduced vendor lock-in. Every acquisition plan needs to prioritize data ownership, interoperability, and privacy. Additionally, agencies must manage rights and safety impacts in AI, particularly concerning biometrics and generative AI, with a full compliance deadline of March 2025. Key Principles and Requirements * An emphasis on privacy, security, data ownership, and system compatibility is the basis of the memorandum’s risk management requirements. A proactive approach ensures agencies address AI’s unique impacts, particularly systems influencing individual rights or safety. Transparency in AI is a central focus. * Agencies are encouraged to work across departments, gathering insights from varied fields to stay aligned with evolving technology. This collaboration supports the adaptation of federal acquisition standards, helping agencies avoid outdated practices and integrate technological advances smoothly. * The memo aims to foster a competitive AI market by introducing shorter contracts, performance-based requirements, and vendor trials. Through these strategies, agencies should retain access to innovative AI solutions without long-term commitments, ensuring a more flexible, dynamic marketplace. For generative AI, the memo sets additional requirements. * The memo recommends using markers such as watermarks or metadata to ensure AI-generated content like synthetic media is clearly distinguishable from genuine media. This control helps prevent confusion and potential misinformation. * Documentation detailing training processes, data sources, and model evaluation is required. Such records enable agencies to verify compliance with federal transparency standards, further securing responsible AI use. * Contract terms must block AI use conflicting with legal and ethical guidelines. Through safeguards, agencies work with vendors to prevent unauthorized or inappropriate applications of AI. For AI providers, preparing for federal acquisitions under M-24-18 involves adopting practical strategies: * Using ISO 42001 standards, vendors can align operations with M-24-18 by establishing ethical, secure AI practices. This framework offers a structured, transparent approach to risk and performance management. * Although not required, FedRAMP certification can improve credibility for cloud-based AI providers, enhancing trust and streamlining the acquisition process by assuring agencies of a provider’s commitment to security. #cybersecurity #AI #ISO #risk

Forvis Mazars US provides details on new DOJ updates calling for corporations to have a mitigation plan to address potential AI risks. Review key takeaways and how they may affect your organization: https://lnkd.in/eZQ2m7Ca

Forvis Mazars US provides details on new DOJ updates calling for corporations to have a mitigation plan to address potential AI risks. Review key takeaways and how they may affect your organization: https://lnkd.in/gkNrFBH3