[SCITT] Re: IETF 120 Topics of Interest

[John Andersen <[e-mail]>]

Most excellent! Please let us know if there's any sample code we can play with.

The stuff I have is rough but it's over here:
- https://github.com/pdxjohnny/scitt-api-emulator/blob/policy_engine_cwt_rebase/scitt_emulator/policy_engine.py
  - LiteLLM sending transparent statement URNs to relying party phase
0 to get workload identity token for tool call:
https://github.com/pdxjohnny/litellm/commit/3b6b7427b15c0cadd23a8b5da639e22a2fba5043#diff-95a105c80600ebb2de10fddc4f64a9754c12815ad7f00e3265400a789ed18786R87-R299

Thank you,
John

On Thu, Jul 11, 2024 at 8:02 AM Michael Prorock <[e-mail]> wrote:
>
> 100%
> we have already run some tests on using SCITT notaries for tracking AI data inputs as well as agent decision making and model components.
>
> I suspect that this will get more formalized and documented publicly as we roll some of those features more broadly out, but that approach is proving valuable for certain key use cases so far.
>
> Mike Prorock
> founder - mesur.io
>
> On Thu, Jul 11, 2024, 09:52 John Andersen <[e-mail]> wrote:
>>
>> Thank you for sharing Orie!
>>
>> Maybe AI agent workload identity would be useful to determine access as well. SCITT could be helpful in keeping an audit trail as AI workloads mutate. Statements around the state of the BOM of the workload could be used to give it access based in its current state.
>>
>> Thank you,
>> John
>>
>> On Thu, Jul 11, 2024 at 08:46 Orie Steele <[e-mail]> wrote:
>>>
>>> Hello,
>>>
>>> I wanted to share some work happening in the IETF that may be of interest to contributors to this list:
>>>
>>> ## IAB Workshop on AI-CONTROL
>>>
>>> Large Language Models and other machine learning techniques require voluminous input data, and one common source of such data is the Internet -- usually, "crawling" Web sites for publicly available content, much in the same way that search engines crawl the Web. This similarity has led to an emerging practice of allowing the Robots Exclusion Protocol (RFC 9309) to control the behavior of AI-oriented crawlers.
>>> ...
>>> - https://mailarchive.ietf.org/arch/msg/ietf-announce/VQvsuYJV6M8gEpm86TTk-LT1CKM/
>>>
>>> ## DIEM Birds of a Feather
>>>
>>> Digital Emblems BoF will discuss the problem of protecting resources with digital emblems (discoverable signatures, credentials, ... tbd).
>>> There has been significant discussion of the problem space and stakeholders.
>>> I would summarize the discussion as:
>>> There are 2 scenarios, protecting digital resources from attack in regions of conflict (cyber space), in accordance with international laws.
>>> Protecting physical and digital resources in accordance with laws (supply chain / cross border trade / customs clearance).
>>> The BoF will likely focus on stakeholders and interest in these use cases, and I would expect a significant discussion of physical and digital supply chain topics.
>>>
>>> - https://datatracker.ietf.org/wg/diem/about/
>>> - https://mailarchive.ietf.org/arch/msg/diem/4oSkrAAWsY-liKfQ1fyutwMPj90/
>>>
>>> Regards,
>>>
>>> OS
>>>
>>>
>>> --
>>> SCITT mailing list -- [e-mail]
>>> To unsubscribe send an email to [e-mail]
>>
>> --
>> SCITT mailing list -- [e-mail]
>> To unsubscribe send an email to [e-mail]