For more than three years, the IEEE Standards Association working group has been improving the draft standards for purchasing artificial intelligence and automated decision-making systems, IEEE 3119-2025. It is designed to help procurement teams identify and manage risks in high-risk areas. Such systems are used by government entities involved in education, health, employment and many other public sector areas. Last year, the Working Group worked with EU agencies to evaluate components of the draft standards and gather information on user needs and their perceptions of the value of the standards.
At that time, the standard included fiveHelp users develop their bidding and identify, mitigate and monitor hazards associated with high-risk AI systems.
These processes are problem definition, supplier evaluation, solution evaluation, contract negotiation and contract monitoring.
Feedback from EU institutions has led the Working Group to reconsider the process and sequence of several activities. The final draft now includes an additional process: recruitment standards, which is carried out after the problem definition process. The Working Group believes that the additional process addresses the challenges of organizations’ experience in preparing AI-specific tenders, such as the need to add transparent and robust data requirements, combined with issues regarding supplier AI governance maturity.
The EU body also stressed that the need to include recruitment standards preparations provides procurement teams with more opportunities to adapt to their bidding and to understand the technical requirements and issues regarding responsible AI system selection. Leaving room for adjustment is especially important when AI acquisitions occur in emerging and changing regulatory environments.
The location of IEEE 3119 in the standard ecosystem
There are currently several internationally recognized standards AI Management,,,,, Artificial Intelligence EthicsAnd general Software acquisition. From the IEEE Standards Association and International Organization for Standardization For AI design, use and lifecycle management.
So far, there is no internationally accepted, consensus-based standard that focuses on the procurement and operation of AI toolsGuide to Responsible Purchase High-risk AI systems For the public interest.
The IEEE 3119 standard solves this gap. Unlike AI Management Standard ISO 42001 and other certifications related to general AI oversight and risk governance, IEEE’s new standards provide Based on risk, operationmethod Help government agencies adapt Traditional Procurement practice.
The government plays an important role in the responsible deployment of AI. However, market dynamics and inequality of AI expertise between industries and governments can be a barrier to success.
One of the core goals of this standard is to better inform purchasing leaders about the items they are going to buy before buying a high-risk AI purchase. IEEE 3119 defines high-risk AI systems as systems that are important to manufacturing or become manufacturing The result decision May have a significant impactAbout people, groups or society. This definition is similar to the definition used in Colorado 2034 You have the filesThe first US state-level law comprehensively solves high-risk systems.
However, the process of this standard does complement ISO 42001 in many ways. The relationship between the two is shown below.
International standards, usually described as Soft methodused to shape AI development and Encourage international cooperation About its governance.
Artificial intelligence or difficult laws with legally binding rules and obligations is a work ongoing around the world. In the United States, Patchwork The state legislates to govern all aspects of AI, and the method of national AI regulations is dispersionwith different federal agencies implementing their own guidelines.
Europe through the EU You have the filesit begins to manage AI systems based on the risk level when the AI system takes effect.
But the world lacks Hard regulatory method With international scope.
The IEEE 3119-2025 standard does match the existing hard method. Due to its focus on procurement, the standard supports Chapter 3 of the EU AI Act and AI behavior in Colorado. The standard also meets the proposed Texas HB 1709 Legislation will require reporting on the use of AI systems for certain commercial entities and state agencies.
because Most AI systems used in the public sector IEEE 3119 is purchased rather than built in-house and is suitable for unwanted commercial AI products and services Substantive modifications Or customize.
Standard target audience
The purpose of this standard is:
- Intermediate procurement professionals and interdisciplinary team members have moderate knowledge of AI governance and AI systems.
- A public and private sector procurement professional who serves as a coordinator or buyer or has an equal role in its body.
- A non-promotion manager and supervisor who is responsible for purchasing or overseeing employees who provide procurement functions.
- Professionals employed by entities involved in public education, utilities, transportation and other publicly funded services that may engage in or manage procurement and are interested in adapting to the purchase process of AI tools.
- AI vendors seeking to understand new transparency and disclosure requirements for their high-risk commercial products and solutions.
Training Program
this IEEE Standards Association and AI Procurement Laboratory supply IEEE is responsible for the training program for artificial intelligence procurement. The course covers how to apply the core process of standard and adapt current practice to the procurement of high-risk AI.
The standard includes over 26 tools and columns in six processes, and the training program explains how many of these tools are used. For example, training includes instructions on how to perform risk Apple analysis, applying vendor evaluation guides for analysis AI vendor claimsand create a “risk register” of AI procurement related to identified use case risks and their potential mitigation. Training courses are now available.
AI integration is still an early days. Policymakers don’t have much experience in buying AI in high-risk areas and mitigating these risks. The IEEE 3119-2025 standard is designed to support institutions to establish and strengthen their AI risk mitigation muscles.
From your website article
Related articles around the Internet
