By co-chairs: Atul Tulshibagwale, SGNL, Tobin South, WorkOS and Stanford, and Jeff Lombardo, AWS

At the recently concluded Identiverse conference, prompted by Jeff Lombardo of AWS, a few of the attendees got together to discuss various security proposals relating to AI. 

It was clear that AI and identity communities have been developing in silos. Such disconnect risks repeating known pitfalls around privacy, security, and interoperability, and may delay the adoption of critical standards that could otherwise accelerate safe and trusted AI deployments. Existing standards only partially cover the emerging needs of AI agents, particularly around delegated authority, agent authentication, propagation and delegation of authorization between agents, and agent discovery and governance.

We realized that we needed a proper forum to discuss the rapidly changing landscape where existing standards, like OAuth 2.0, can help accelerate the AI community’s objectives, what use cases are critical paths, and how to close gaps in the current de facto standard and standards landscape standing at the intersection of AI and identity. 

The idea of an OpenID Foundation Community Group was born there, not with the intent to control the narrative in a single standards body, but as a credible and open forum to bring conversations already happening and triage where to progress them further. To that end, it was agreed that discussions and deliverables from these conversations should lead to a shared roadmap and deliverables that progress in AI and identity forums, and within de facto standards efforts as well as open standards bodies like OIDF and its peers. 

Together we hope to accelerate the delivery of critical paths and a consensus-based roadmap that helps accelerate the AI community, while leveraging the expertise of identity experts, averting known pitfalls and finding the shared path to ‘safe pastures.’

To say that there are a huge number of AI security proposals that touch upon identity in some way, is an understatement. A place where interested parties can freely discuss these proposals and ideas without having to worry about intellectual property concerns has, however, been missing - until now. Recognizing this, the OpenID Foundation Board approved the creation of a community group, named ‘Artificial Intelligence Identity Management’ (AIIM) following the OIDF Community Group policy.

What is a Community Group?

A Working Group (WG) in the OpenID Foundation tackles specific problem areas and delivers an output in the form of a specification, or series of specifications or documents. Community Groups (CGs) in the OpenID Foundation, on the other hand, are explicitly designed not to produce specifications, but to create a safe space to convene. 

To progress discussions that can identify ecosystem gaps, adoption gaps, or that operate across working groups where all discussions and deliverables are protected by a mutually agreed CG Charter and Participation Agreement. If gaps in standards are identified they are brought to the relevant OIDF WG or liaison partner standards body. Because of the participation agreement requirement, participants can freely discuss ideas with work products protected by OIDF Intellectual Property so the work is freely available to all. You can find the website for the website for the AIIM Community Group here with the link to the calls 9am Pacific Time Thursdays, and the Participation Agreement for the new AIIM CG here.

There is no fee to participate in an OIDF CG, reducing the barriers for AI and identity experts to take part in the CG conversations.  

Aims of the AIIM CG

“AI is disrupting many dimensions of the internet…” So begins the charter of the AIIM CG.

This disruption spans social interactions, digital commerce, financial services, and the broader human/digital interface. However, as AI systems scale rapidly, a critical challenge is emerging. The AI and identity communities are largely developing in silos. This disconnect risks repeating known pitfalls around privacy, security, and interoperability, and may delay the adoption of critical standards that could otherwise accelerate safe and trusted AI deployments.

We will work to:

• Identify areas that are not currently addressed by standards, but which need to be addressed and where the identity community needs to focus
• Arrive at a consensus on terminology
• Engage with the industry stakeholders, including major platform vendors
• Define ‘Agentic AI champion use cases’ that trusted partner organizations can refer to or leverage in their work
• Monitor government regulations regarding AI that impact identity

And work on an AI white paper, which the co-chairs are committed to authoring.

The AIIM CG will operate under the core principles of respect, privacy through consent, and interoperability, aiming to support scalable, inclusive, and trusted AI solutions. While it will not develop standards protocols directly, it will lay the essential groundwork for future standards development within OIDF or through liaison partnerships.

Proposers of the AIIM CG

Many thought leaders in the first Identiverse meeting and those working on the OIDF AI and Identity whitepaper have already signed up under their own initiative to support this CG formation. No commitment by their current employer should be assumed based on the proposed support by the individuals. The proposers include:   

• Gail Hodges  
• Atul Tulshibagwale 
• George Fletcher  
• Jeff Lombardo 
• Tobin South 
• Naveen CM 
• Aaron Parecki  
• Sean O’Dell  
• Nancy Cam-Winget  
• Mike Kiser 
• Alexandre Babeanu  

Why you should participate

If you need a safe space to discuss ideas related to AI and identity, this is it. Since the pace of development here is furious, we expect there to be a lot of noise, and as co-chairs, our job is to make sure we don’t lose focus on what really matters to the identity community. Join the AIIM CG by signing the participation agreement linked to here.

We welcome everyone to join the regular weekly calls on Thursdays at 9 AM PT. See the AIIM community group home page to get the details. You can also subscribe to the AIIM CG mail list, and if you are a contributor, post to it to discuss ideas. The OpenID Slack also has a channel for the AIIM CG.

Looking forward to amazing discussions here!