rfc9969.original.xml   rfc9969.xml 
<?xml version='1.0' encoding='utf-8'?> <?xml version='1.0' encoding='utf-8'?>
<!DOCTYPE rfc [ <!DOCTYPE rfc [
<!ENTITY nbsp "&#160;"> <!ENTITY nbsp "&#160;">
<!ENTITY zwsp "&#8203;"> <!ENTITY zwsp "&#8203;">
<!ENTITY nbhy "&#8209;"> <!ENTITY nbhy "&#8209;">
<!ENTITY wj "&#8288;"> <!ENTITY wj "&#8288;">
]> ]>
<?xml-stylesheet type="text/xsl" href="rfc2629.xslt" ?> <rfc xmlns:xi="http://www.w3.org/2001/XInclude" ipr="trust200902" docName="draft
<!-- generated by https://github.com/cabo/kramdown-rfc version 1.7.29 (Ruby 3.4. -iab-ai-control-report-02" category="info" consensus="true" submissionType="IAB"
4) --> xml:lang="en" number="9969" tocInclude="true" sortRefs="true" symRefs="true" ve
<?rfc compact="yes"?> rsion="3">
<?rfc subcompact="yes"?> <link href="https://datatracker.ietf.org/doc/draft-iab-ai-control-report-02" r
<rfc xmlns:xi="http://www.w3.org/2001/XInclude" ipr="trust200902" docName="draft el="prev"/>
-iab-ai-control-report-02" category="info" consensus="true" submissionType="IAB"
version="3">
<!-- xml2rfc v2v3 conversion 3.30.1 -->
<front> <front>
<title>IAB AI-CONTROL Workshop Report</title> <title abbrev="IAB AI-CONTROL Workshop Report">Report from the IAB Workshop
<seriesInfo name="Internet-Draft" value="draft-iab-ai-control-report-02"/> on AI-CONTROL</title>
<seriesInfo name="RFC" value="9969"/>
<author initials="M." surname="Nottingham" fullname="Mark Nottingham"> <author initials="M." surname="Nottingham" fullname="Mark Nottingham">
<organization/> <organization/>
<address> <address>
<postal> <postal>
<postalLine>Melbourne</postalLine> <city>Melbourne</city>
<postalLine>Australia</postalLine> <country>Australia</country>
</postal> </postal>
<email>mnot@mnot.net</email> <email>mnot@mnot.net</email>
<uri>https://www.mnot.net/</uri> <uri>https://www.mnot.net/</uri>
</address> </address>
</author> </author>
<author initials="S." surname="Krishnan" fullname="Suresh Krishnan"> <author initials="S." surname="Krishnan" fullname="Suresh Krishnan">
<organization/> <organization/>
<address> <address>
<email>suresh.krishnan@gmail.com</email> <email>suresh.krishnan@gmail.com</email>
</address> </address>
</author> </author>
<date year="2025" month="September" day="06"/> <date year="2026" month="May"/>
<keyword>policy</keyword> <keyword>policy</keyword>
<keyword>Artificial Intelligence</keyword> <keyword>Artificial Intelligence</keyword>
<keyword>Robots Exclusion Protocol</keyword> <keyword>Robots Exclusion Protocol</keyword>
<keyword>web crawler</keyword> <keyword>web crawler</keyword>
<keyword>robots.txt</keyword> <keyword>robots.txt</keyword>
<abstract> <abstract>
<?line 87?> <?line 91?>
<t>The AI-CONTROL Workshop was convened by the Internet Architecture Board (IAB) in September 2024. This report summarizes its significant points of discussion and identifies topics that may warrant further consideration and work.</t> <t>The AI-CONTROL Workshop was convened by the Internet Architecture Board (IAB) in September 2024. This report summarizes its significant points of discussion and identifies topics that may warrant further consideration and work.</t>
<t>Note that this document is a report on the proceedings of the workshop. The views and positions documented in this report are those of the workshop pa rticipants and do not necessarily reflect IAB views and positions.</t> <t>Note that this document is a report on the proceedings of the workshop. The views and positions documented in this report are those of the workshop pa rticipants and do not necessarily reflect IAB views and positions.</t>
</abstract> </abstract>
<note removeInRFC="true">
<name>Discussion Venues</name>
<t>Source for this draft and an issue tracker can be found at
<eref target="https://github.com/intarchboard/draft-iab-ai-control-report"/>
.</t>
</note>
</front> </front>
<middle> <middle>
<?line 93?> <?line 97?>
<section anchor="introduction"> <section anchor="introduction">
<name>Introduction</name> <name>Introduction</name>
<t>The Internet Architecture Board (IAB) holds occasional workshops design ed to consider long-term issues and strategies for the Internet, and to suggest future directions for the Internet architecture. This long-term planning functio n of the IAB is complementary to the ongoing engineering efforts performed by wo rking groups of the Internet Engineering Task Force (IETF).</t> <t>The Internet Architecture Board (IAB) holds occasional workshops design ed to consider long-term issues and strategies for the Internet, and to suggest future directions for the Internet architecture. This long-term planning functio n of the IAB is complementary to the ongoing engineering efforts performed by wo rking groups of the Internet Engineering Task Force (IETF).</t>
<t>The Internet is one of the major sources of data used to train large la nguage models (Large Language Models (LLMs), or more generally, "Artificial Inte lligence (AI)"). Because this use was not envisioned by most publishers of infor mation on the Internet, a means of expressing the owners' preferences regarding AI crawling has emerged, sometimes backed by law (e.g., in the European Union's AI Act <xref target="AI-ACT"/>).</t> <t>The Internet is one of the major sources of data used to train Large La nguage Models (LLMs) (or, more generally, Artificial Intelligence (AI)). Because this use was not envisioned by most publishers of information on the Internet, a means of expressing the owners' preferences regarding AI crawling has emerged, sometimes backed by law (e.g., in the European Union's AI Act <xref target="AI- ACT"/>).</t>
<t>The IAB convened the AI-CONTROL Workshop on 19-20 September 2024 to "ex plore practical opt-out mechanisms for AI and build an understanding of use case s, requirements, and other considerations in this space" <xref target="CFP"/>. I n particular, the emerging practice of using the Robots Exclusion Protocol <xref target="RFC9309"/> -- also known as "robots.txt" -- has not been coordinated be tween AI crawlers, resulting in considerable differences in how they treat it. F urthermore, robots.txt may or may not be a suitable way to control AI crawlers. However, discussion was not limited to consideration of robots.txt, and approach es other than opt-out were considered.</t> <t>The IAB convened the AI-CONTROL Workshop on 19-20 September 2024 to "ex plore practical opt-out mechanisms for AI and build an understanding of use case s, requirements, and other considerations in this space" <xref target="CFP"/>. I n particular, the emerging practice of using the Robots Exclusion Protocol <xref target="RFC9309"/> -- also known as "robots.txt" -- has not been coordinated be tween AI crawlers, resulting in considerable differences in how they treat it. F urthermore, robots.txt may or may not be a suitable way to control AI crawlers. However, discussion was not limited to consideration of robots.txt, and approach es other than opt-out were considered.</t>
<t>To ensure many viewpoints were represented, the program committee invit ed a broad selection of technical experts, AI vendors, content publishers, civil society advocates, and policymakers.</t> <t>To ensure many viewpoints were represented, the program committee invit ed a broad selection of technical experts, AI vendors, content publishers, civil society advocates, and policymakers.</t>
<section anchor="chatham-house-rule"> <section anchor="chatham-house-rule">
<name>Chatham House Rule</name> <name>Chatham House Rule</name>
<t>Participants agreed to conduct the workshop under the Chatham House R ule <xref target="CHATHAM-HOUSE"/>, so this report does not attribute statements to individuals or organizations without express permission. Most submissions to the workshop were public and thus attributable; they are used here to provide s ubstance and context.</t> <t>Participants agreed to conduct the workshop under the Chatham House R ule <xref target="CHATHAM-HOUSE"/>, so this report does not attribute statements to individuals or organizations without express permission. Most submissions to the workshop were public and thus attributable; they are used here to provide s ubstance and context.</t>
<t><xref target="attendees"/> lists the workshop participants, unless th ey requested that this information be withheld.</t> <t><xref target="attendees"/> lists the workshop participants, unless th ey requested that this information be withheld.</t>
</section> </section>
<section anchor="views-expressed-in-this-report"> <section anchor="views-expressed-in-this-report">
<name>Views Expressed in this Report</name> <name>Views Expressed in This Report</name>
<t>This document is a report on the proceedings of the workshop. The vie <t>This document is a report on the proceedings of the workshop. The vie
ws and positions documented in this report are expressed during the workshop by ws and positions documented in this report are expressed during the workshop by
participants and do not necessarily reflect IAB's views and positions.</t> participants and do not necessarily reflect the IAB's views and positions.</t>
<t>Furthermore, the content of the report comes from presentations given <t>Furthermore, the content of the report comes from presentations given
by workshop participants and notes taken during the discussions, without interp by workshop participants and notes taken during the discussions, without interp
retation or validation. Thus, the content of this report follows the flow and di retation or validation. Thus, the content of this report follows the flow and di
alogue of the workshop but does not attempt to capture a consensus.</t> alog of the workshop but does not attempt to capture a consensus.</t>
</section> </section>
</section> </section>
<section anchor="workshop-scope-and-discussion"> <section anchor="workshop-scope-and-discussion">
<name>Workshop Scope and Discussion</name> <name>Workshop Scope and Discussion</name>
<t>The workshop began by surveying the state of AI control.</t> <t>The workshop began by surveying the state of AI control.</t>
<t>Currently, Internet publishers express their preferences for how their content is treated for purposes of AI training using a variety of mechanisms, in cluding declarative ones, such as terms of service, embedded metadata, and robot s.txt <xref target="RFC9309"/>, and active ones, such as use of paywalls and sel ective blocking of crawlers (e.g., by IP address, User-Agent).</t> <t>Currently, Internet publishers express their preferences for how their content is treated for the purposes of AI training using a variety of mechanisms . These include declarative mechanisms, such as terms of service, embedded metad ata, and robots.txt <xref target="RFC9309"/>, as well as active mechanisms, such as use of paywalls and selective blocking of crawlers (e.g., by IP address or U ser-Agent).</t>
<t>There was disagreement about the implications of AI opt-out overall. Re search presented at the workshop <xref target="DECLINE"/> indicates that the use of such controls is becoming more prevalent, reducing the availability of data to AI (for purposes including training and inference-time usage). Some of the pa rticipants expressed concern about the implications of this -- although at least one AI vendor seemed less concerned by this, indicating that "there are plenty of tokens available" for training, even if many opt out. Others expressed a need to opt out of AI training because of how they perceive its effects on their con trol over content, seeing AI as usurping their relationships with customers and a potential threat to whole industries.</t> <t>There was disagreement about the implications of AI opt-out overall. Re search presented at the workshop <xref target="DECLINE"/> indicates that the use of such controls is becoming more prevalent, reducing the availability of data to AI (for purposes including training and inference-time usage). Some of the pa rticipants expressed concern about the implications of this -- although at least one AI vendor seemed less concerned by this, indicating that "there are plenty of tokens available" for training, even if many opt out. Others expressed a need to opt out of AI training because of how they perceive its effects on their con trol over content, seeing AI as usurping their relationships with customers and a potential threat to whole industries.</t>
<t>However, there was quick agreement that both viewpoints were harmed by the current state of AI opt-out -- a situation where "no one is better off" (in the words of one participant).</t> <t>However, there was quick agreement that both viewpoints were harmed by the current state of AI opt-out -- a situation where "no one is better off" (in the words of one participant).</t>
<t>Much of that dysfunction was attributed to the lack of coordination and standards for AI opt-out. Currently, content publishers need to consult with ea ch AI vendor to understand how to opt out of training their products, as there i s significant variance in each vendor's behaviour. Furthermore, publishers need to continually monitor both for new vendors, and for changes to the policies of the vendors they are aware of.</t> <t>Much of that dysfunction was attributed to the lack of coordination and standards for AI opt-out. Currently, content publishers need to consult with ea ch AI vendor to understand how to opt out of training their products, as there i s significant variance in each vendor's behavior. Furthermore, publishers need t o continually monitor both new vendors and policy updates from the vendors they are aware of.</t>
<t>Underlying those immediate issues, however, are significant constraints that could be attributed to uncertainties in the legal context, the nature of A I, and the implications of needing to opt out of crawling for it.</t> <t>Underlying those immediate issues, however, are significant constraints that could be attributed to uncertainties in the legal context, the nature of A I, and the implications of needing to opt out of crawling for it.</t>
<section anchor="crawl-time-vs-inference-time"> <section anchor="crawl-time-vs-inference-time">
<name>Crawl Time vs. Inference Time</name> <name>Crawl Time vs. Inference Time</name>
<t>Perhaps most significant is the "crawl time vs. inference time" probl em. Statements of preference are apparent at crawl time, bound to content either by location (e.g., robots.txt) or embedded inside the content itself as metadat a. However, the target of those directives is often disassociated from the crawl er, either because the crawl data is not only used for training AI models, or be cause the preferences could be applicable at inference time.</t> <t>Perhaps most significant is the "crawl time vs. inference time" probl em. Statements of preference are apparent at crawl time, bound to content either by location (e.g., robots.txt) or embedded inside the content itself as metadat a. However, the target of those directives is often disassociated from the crawl er, either because the crawl data is not only used for training AI models or bec ause the preferences could be applicable at inference time.</t>
<section anchor="multiple-uses-for-crawl-data"> <section anchor="multiple-uses-for-crawl-data">
<name>Multiple Uses for Crawl Data</name> <name>Multiple Uses for Crawl Data</name>
<t>A crawl's data might have multiple uses because the vendor also has <t>A crawl's data might have multiple uses because the vendor also has
another product that uses it (e.g., a search engine), or because the crawl is p another product that uses it (e.g., a search engine) or because the crawl is pe
erformed by a party other than the AI vendor. Both are very common patterns: ope rformed by a party other than the AI vendor. Both are very common patterns: Oper
rators of many Internet search engines also train AI models, and many AI models ators of many Internet search engines also train AI models, and many AI models u
use third-party crawl data. In either case, conflating different uses can change se third-party crawl data. In either case, conflating different uses can change
the incentives for publishers to cooperate with the crawler.</t> the incentives for publishers to cooperate with the crawler.</t>
<t>Well-established uses of crawling, such as Internet search, were se <t>Well-established uses of crawling, such as Internet searches, were
en by participants as at least partially aligned with the interests of publisher seen by participants as at least partially aligned with the interests of publish
s: they allow their sites to be crawled, and in return, they receive higher traf ers: They allow their sites to be crawled, and in return, they receive higher tr
fic and attention due to being in the search index. However, several participant affic and attention due to being in the search index. However, several participa
s pointed out that this symbiotic relationship does not exist for AI training us nts pointed out that this symbiotic relationship does not exist for AI training
es -- with some viewing AI as hostile to publishers, because it has the capacity uses -- with some viewing AI as hostile to publishers because it has the capacit
to take traffic away from their sites.</t> y to take traffic away from their sites.</t>
<t>Therefore, when a crawler has multiple uses that include AI, partic ipants observed that "collateral damage" was likely for non-AI uses, especially when publishers take more active control measures, such as blocking or paywalls, to protect their interests.</t> <t>Therefore, when a crawler has multiple uses that include AI, partic ipants observed that "collateral damage" was likely for non-AI uses, especially when publishers take more active control measures, such as blocking or paywalls, to protect their interests.</t>
<t>Several participants expressed concerns about this phenomenon's eff ects on the ecosystem, effectively "locking down the Web" with one opining that there were implications for freedom of expression overall.</t> <t>Several participants expressed concerns about this phenomenon's eff ects on the ecosystem, effectively "locking down the Web" with one opining that there were implications for freedom of expression overall.</t>
</section> </section>
<section anchor="application-of-preferences"> <section anchor="application-of-preferences">
<name>Application of Preferences</name> <name>Application of Preferences</name>
<t>When data is used to train an LLM, the resulting model does not hav <t>When data is used to train an LLM, the resulting model does not hav
e the ability to only selectively use a portion of it when performing a task, be e the ability to only selectively use a portion of it when performing a task bec
cause inference uses the whole model, and it is not possible to identify specifi ause inference uses the whole model, and it is not possible to identify specific
c input data for its use in doing so.</t> input data for its use in doing so.</t>
<t>This means that while publishers' preferences may be available when <t>This means that while publishers' preferences may be available when
content is crawled, they generally are not when inference takes place. Those pr content is crawled, they generally are not when inference takes place. Those pr
eferences that are stated in reference to use by AI -- for example, "no military eferences that are stated in reference to use by AI -- for example, "no military
uses" or "non-commercial only" cannot be applied by a general-purpose "foundati uses" or "non-commercial only" -- cannot be applied by a general-purpose "found
on" model.</t> ation" model.</t>
<t>This leaves a few unappealing choices to AI vendors that wish to co <t>This leaves a few unappealing choices to AI vendors that wish to co
mply with those preferences. They can simply omit such data from foundation mode mply with those preferences. They can simply omit such data from foundation mode
ls, thereby reducing their viability. Or, they can create a separate model for e ls, thereby reducing their viability. Or they can create a separate model for ea
ach permutation of preferences -- with a likely proliferation of models as the s ch permutation of preferences -- with a likely proliferation of models as the se
et of permutations expands.</t> t of permutations expands.</t>
<t>Compounding this issue was the observation that preferences change over time, whereas LLMs are created over long time frames and cannot easily be u pdated to reflect those changes. Of particular concern to some was how this make s an opt-out regime "stickier" because content that has no associated preference (such as that which predates the authors' knowledge of LLMs) is allowed to be u sed for these unforeseen purposes.</t> <t>Compounding this issue was the observation that preferences change over time, whereas LLMs are created over long time frames and cannot easily be u pdated to reflect those changes. Of particular concern to some was how this make s an opt-out regime "stickier" because content that has no associated preference (such as that which predates the authors' knowledge of LLMs) is allowed to be u sed for these unforeseen purposes.</t>
</section> </section>
</section> </section>
<section anchor="trust"> <section anchor="trust">
<name>Trust</name> <name>Trust</name>
<t>This disconnection between the statement of preferences and its appli cation was felt by participants to contribute to a lack of trust in the ecosyste m, along with the typical lack of attribution for data sources in LLMs, lack of an incentive for publishers to contribute data, and finally (and most noted) a l ack of any means of monitoring compliance with preferences.</t> <t>Participants felt that the disconnection between the statement of pre ferences and its application contribute to a lack of trust in the ecosystem, alo ng with the typical lack of attribution for data sources in LLMs, a lack of an i ncentive for publishers to contribute data, and finally (and most noted) a lack of any means of monitoring compliance with preferences.</t>
<t>This lack of trust led some participants to question whether communic ating preferences is sufficient in all cases without an accompanying way to enfo rce them, or even to audit adherence to them. Some participants also indicated t hat a lack of trust was the primary cause of the increasingly prevalent blocking of AI crawler IP addresses, among other measures.</t> <t>This lack of trust led some participants to question whether communic ating preferences is sufficient in all cases without an accompanying way to enfo rce them, or even to audit adherence to them. Some participants also indicated t hat a lack of trust was the primary cause of the increasingly prevalent blocking of AI crawler IP addresses, among other measures.</t>
</section> </section>
<section anchor="attachment"> <section anchor="attachment">
<name>Attachment</name> <name>Attachment</name>
<t>One of the primary focuses of the workshop was on <em>attachment</em> -- how preferences are associated with content on the Internet. A range of mech anisms was discussed.</t> <t>One of the primary focuses of the workshop was on <em>attachment</em> , i.e., how preferences are associated with content on the Internet. A range of mechanisms was discussed.</t>
<section anchor="robotstxt-and-similar"> <section anchor="robotstxt-and-similar">
<name>robots.txt (and similar)</name> <name>robots.txt (and Similar)</name>
<t>The Robots Exclusion Protocol <xref target="RFC9309"/> is widely re <t>The Robots Exclusion Protocol <xref target="RFC9309"/> is widely re
cognised by AI vendors as an attachment mechanism for preferences. Several defic cognized by AI vendors as an attachment mechanism for preferences. Several defic
iencies were discussed.</t> iencies were discussed.</t>
<t>First, it does not scale to offer granular control over large sites where authors might want to express different policies for a range of content ( for example, YouTube).</t> <t>First, it does not scale to offer granular control over large sites where authors might want to express different policies for a range of content ( for example, YouTube).</t>
<t>Robots.txt is also typically under the control of the site administ <t>robots.txt is also typically under the control of the site administ
rator. If a site has content from many creators (as is often the case for social rator. If a site has content from many creators (as is often the case for social
media and similar platforms), the administrator may not allow them to express t media and similar platforms), the administrator may not allow them to express t
heir preferences fully, or at all.</t> heir preferences fully, or at all.</t>
<t>If content is copied or moved to a different site, the preferences <t>If content is copied or moved to a different site, the preferences
at the new site need to be explicitly transferred, because robots.txt is a separ at the new site need to be explicitly transferred because robots.txt is a separa
ate resource.</t> te resource.</t>
<t>These deficiencies led many participants to feel that robots.txt ca <t>These deficiencies led many participants to feel that robots.txt ca
nnot be the only solution to opt-out: rather, it should be part of a larger syst nnot be the only solution to opt-out: Rather, it should be part of a larger syst
em that addresses its shortcomings.</t> em that addresses its shortcomings.</t>
<t>Participants noted that other, similar attachment mechanisms have b <t>Participants noted that other similar attachment mechanisms have be
een proposed. However, none appear to have gained as much attention or implement en proposed. However, none appear to have gained as much attention or implementa
ation (both by AI vendors and content owners) as robots.txt.</t> tion (both by AI vendors and content owners) as robots.txt.</t>
</section> </section>
<section anchor="embedding"> <section anchor="embedding">
<name>Embedding</name> <name>Embedding</name>
<t>Another mechanism for associating preferences with content is to em <t>Another mechanism for associating preferences with content is to em
bed them into the content itself. Many formats used on the Internet allow this; bed them into the content itself. Many formats used on the Internet allow this;
for example, HTML has the <tt>&lt;meta&gt;</tt> tag, images have XMP and similar for example, HTML has the <tt>&lt;meta&gt;</tt> tag, images have Extensible Meta
metadata sections, and XML and JSON have rich potential for extensions to carry data Platform (XMP) and similar metadata sections, and XML and JSON have rich po
such data.</t> tential for extensions to carry such data.</t>
<t>Embedded preferences were seen to have the advantage of granularity <t>Embedded preferences were seen to have the advantage of granularity
, and of "travelling with" content as it is produced, when it is moved from site , and of "traveling with" content as it is produced, when the content that embed
to site, or when it is stored offline.</t> s the preferences is moved from site to site or when it is stored offline.</t>
<t>However, several participants pointed out that embedded preferences <t>However, several participants pointed out that embedded preferences
are easily stripped from most formats. This is a common practice for reducing t are easily stripped from most formats. This is a common practice for reducing t
he size of a file (thereby improving performance when downloading it), and for a he size of a file (thereby improving performance when downloading it) and for as
ssuring privacy (since metadata often leaks information unintentionally).</t> suring privacy (since metadata often leaks information unintentionally).</t>
<t>Furthermore, some types of content are not suitable for embedding. <t>Furthermore, some types of content are not suitable for embedding.
For example, it is not possible to embed preferences into purely textual content For example, it is not possible to embed preferences into purely textual content
, and Web pages with content from several producers (such as a social media or c , and web pages with content from several producers (such as a social media or c
omments feed) cannot easily reflect preferences for each one.</t> omment feeds) cannot easily reflect preferences for each one.</t>
<t>Participants noted that the means of embedding preferences in many <t>Participants noted that the means of embedding preferences in many
formats would need to be determined by or coordinated with organisations outside formats would need to be determined by or coordinated with organizations outside
the IETF. For example, HTML and many image formats are maintained by external b the IETF. For example, HTML and many image formats are maintained by external b
odies.</t> odies.</t>
</section> </section>
<section anchor="registries"> <section anchor="registries">
<name>Registries</name> <name>Registries</name>
<t>In some existing copyright management regimes, it is already common to have a registry of works that is consulted upon use. For example, this appro ach is often used for photographs, music, and video.</t> <t>In some existing copyright management regimes, it is already common to have a registry of works that is consulted upon use. For example, this appro ach is often used for photographs, music, and video.</t>
<t>Typically, registries use hashing mechanisms to create a "fingerpri nt" for the content that is robust to changes.</t> <t>Typically, registries use hashing mechanisms to create a "fingerpri nt" for the content that is robust to changes.</t>
<t>Using a registry decouples the content in question from its locatio <t>Using a registry decouples the content in question from its locatio
n, so that it can be found even if moved. It is also claimed to be robust agains n so that it can be found even if moved. It is also claimed to be robust against
t stripping of embedded metadata, which is a common practice to improve performa stripping of embedded metadata, which is a common practice to improve performan
nce and/or privacy.</t> ce and/or privacy.</t>
<t>However, several participants pointed out issues with deploying reg <t>However, several participants pointed out issues with deploying reg
istries at Internet scale. While they may be effective for (relatively) closed a istries at the scale of the Internet. While they may be effective for (relativel
nd well-known ecosystems such as commercial music publishing, applying them to a y) closed and well-known ecosystems, such as commercial music publishing, applyi
diverse and very large ecosystem like the Internet has proven problematic.</t> ng them to a diverse and very large ecosystem like the Internet has proven probl
ematic.</t>
</section> </section>
</section> </section>
<section anchor="vocabulary"> <section anchor="vocabulary">
<name>Vocabulary</name> <name>Vocabulary</name>
<t>Another major focus area for the workshop was on <em>vocabulary</em> -- the specific semantics of the opt-out signal. Several participants noted that there are already many proposals for vocabularies, as well as many conflicting vocabularies already in use. Several examples were discussed, including where ex isting terms were ambiguous, did not address common use cases, or were used in c onflicting ways by different actors.</t> <t>Another major focus area for the workshop was on <em>vocabulary</em> -- the specific semantics of the opt-out signal. Several participants noted that there are already many proposals for vocabularies, as well as many conflicting vocabularies already in use. Several examples were discussed, including where ex isting terms were ambiguous, did not address common use cases, or were used in c onflicting ways by different actors.</t>
<t>Although no conclusions regarding exact vocabulary were reached, it w as generally agreed that a complex vocabulary is unlikely to succeed.</t> <t>Although no conclusions regarding exact vocabulary were reached, it w as generally agreed that a complex vocabulary is unlikely to succeed.</t>
</section> </section>
</section> </section>
<section anchor="conclusions"> <section anchor="conclusions">
<name>Conclusions</name> <name>Conclusions</name>
<t>Participants generally agreed that on its current path, the ecosystem i s not sustainable. As one remarked, "robots.txt is broken and we broke it."</t> <t>Participants generally agreed that on its current path, the ecosystem i s not sustainable. As one remarked, "robots.txt is broken and we broke it".</t>
<t>Legal uncertainty, along with fundamental limitations of opt-out regime s pointed out above, limit the effectiveness of any technical solution, which wi ll be operating in a system unlike either robots.txt (where there is a symbiotic relationship between content owners and the crawlers) or copyright (where the d efault is effectively opt-in, not opt-out).</t> <t>Legal uncertainty, along with fundamental limitations of opt-out regime s pointed out above, limit the effectiveness of any technical solution, which wi ll be operating in a system unlike either robots.txt (where there is a symbiotic relationship between content owners and the crawlers) or copyright (where the d efault is effectively opt-in, not opt-out).</t>
<t>However, the workshop ended with general agreement that positive steps could be taken to improve the communication of preferences from content owners f or AI use cases. In discussion, it was evident that the discovery of preferences from multiple attachment mechanisms is necessary to meet the diverse needs of c ontent authors, and that therefore defining how they are combined is important.< /t> <t>However, the workshop ended with general agreement that positive steps could be taken to improve the communication of preferences from content owners f or AI use cases. In discussion, it was evident that the discovery of preferences from multiple attachment mechanisms is necessary to meet the diverse needs of c ontent authors and, therefore, that defining how they are combined is important. </t>
<t>We outline a proposed standard program below.</t> <t>We outline a proposed standard program below.</t>
<section anchor="potential-standards-work"> <section anchor="potential-standards-work">
<name>Potential Standards Work</name> <name>Potential Standards Work</name>
<t>The following items were felt to be good starting points for IETF wor <t>The following items were identified as good starting points for IETF
k:</t> work:</t>
<ul spacing="compact"> <ul spacing="normal">
<li> <li>
<t>Attachment to Web sites by location (in robots.txt or a similar m echanism)</t> <t>Attachment to websites by location (in robots.txt or a similar me chanism)</t>
</li> </li>
<li> <li>
<t>Attachment via embedding in IETF-controlled formats (e.g., HTTP h eaders)</t> <t>Attachment via embedding in IETF-controlled formats (e.g., HTTP h eaders)</t>
</li> </li>
<li> <li>
<t>Definition of a common core vocabulary</t> <t>Definition of a common core vocabulary</t>
</li> </li>
<li> <li>
<t>Definition of the overall regime; e.g., how to combine preference s discovered from multiple attachment mechanisms</t> <t>Definition of the overall regime, e.g., how to combine preference s discovered from multiple attachment mechanisms</t>
</li> </li>
</ul> </ul>
<t>It would be expected that the IETF would coordinate with other SDOs t o define embedding in other formats (e.g., HTML).</t> <t>It would be expected that the IETF would coordinate with other Standa rds Development Organizations (SDOs) to define embedding in other formats (e.g., HTML).</t>
<section anchor="out-of-initial-scope"> <section anchor="out-of-initial-scope">
<name>Out of Initial Scope</name> <name>Out of Initial Scope</name>
<t>It was broadly agreed that it would not be useful to work on the fo llowing items, at least to begin with:</t> <t>It was broadly agreed that it would not be useful to work on the fo llowing items, at least to begin with:</t>
<ul spacing="compact"> <ul spacing="normal">
<li> <li>
<t>Enforcement mechanisms for preferences</t> <t>Enforcement mechanisms for preferences</t>
</li> </li>
<li> <li>
<t>Registry-based solutions</t> <t>Registry-based solutions</t>
</li> </li>
<li> <li>
<t>Identifying or authenticating crawlers and/or content owners</t > <t>Identifying or authenticating crawlers and/or content owners</t >
</li> </li>
<li> <li>
<t>Audit or transparency mechanisms</t> <t>Audit or transparency mechanisms</t>
</li> </li>
</ul> </ul>
</section> </section>
</section> </section>
</section> </section>
<section anchor="iana-considerations">
<name>IANA Considerations</name>
<t>This document has no IANA actions.</t>
</section>
<section anchor="security-considerations"> <section anchor="security-considerations">
<name>Security Considerations</name> <name>Security Considerations</name>
<t>This document is a workshop report and does not impact the security of the Internet.</t> <t>This document is a workshop report and does not impact the security of the Internet.</t>
</section> </section>
</middle> </middle>
<back> <back>
<references anchor="sec-informative-references"> <references anchor="sec-informative-references">
<name>Informative References</name> <name>Informative References</name>
<reference anchor="CHATHAM-HOUSE" target="https://www.chathamhouse.org/abo ut-us/chatham-house-rule"> <reference anchor="CHATHAM-HOUSE" target="https://www.chathamhouse.org/abo ut-us/chatham-house-rule">
<front> <front>
<title>Chatham House Rule</title> <title>Chatham House Rule</title>
<author> <author>
<organization>Chatham House</organization> <organization>Chatham House</organization>
</author> </author>
<date>n.d.</date> <date/>
</front> </front>
</reference> </reference>
<reference anchor="CFP" target="https://datatracker.ietf.org/group/aicontr olws/about/"> <reference anchor="CFP" target="https://datatracker.ietf.org/group/aicontr olws/about/">
<front> <front>
<title>IAB Workshop on AI-CONTROL</title> <title>IAB Workshop on AI-CONTROL</title>
<author> <author>
<organization>Internet Architecture Board</organization> <organization>Internet Architecture Board</organization>
</author> </author>
<date>n.d.</date> <date/>
</front> </front>
</reference> </reference>
<reference anchor="PAPERS" target="https://datatracker.ietf.org/group/aico ntrolws/materials/"> <reference anchor="PAPERS" target="https://datatracker.ietf.org/group/aico ntrolws/materials/">
<front> <front>
<title>IAB Workshop on AI-CONTROL Materials</title> <title>IAB Workshop on AI-CONTROL Materials</title>
<author> <author>
<organization>Internet Architecture Board</organization> <organization>Internet Architecture Board</organization>
</author> </author>
<date>n.d.</date> <date/>
</front> </front>
</reference> </reference>
<reference anchor="AI-ACT" target="https://eur-lex.europa.eu/eli/reg/2024/ 1689/oj"> <reference anchor="AI-ACT" target="https://eur-lex.europa.eu/eli/reg/2024/ 1689/oj">
<front> <front>
<title>Regulation (eu) 2024/1689 of the European Parliament and of the Council</title> <title>Regulation (EU) 2024/1689 of the European Parliament and of the Council of 13 June 2024 laying down harmonised rules on artificial intelligence and amending Regulations (EC) No 300/2008, (EU) No 167/2013, (EU) No 168/2013, (EU) 2018/858, (EU) 2018/1139 and (EU) 2019/2144 and Directives 2014/90/EU, (EU) 2016/797 and (EU) 2020/1828 (Artificial Intelligence Act) (Text with EEA releva nce)</title>
<author> <author>
<organization>European Parliament</organization> <organization>European Parliament</organization>
</author> </author>
<date year="2024" month="June" day="13"/> <date year="2024" month="June" day="13"/>
</front> </front>
</reference> </reference>
<reference anchor="DECLINE" target="https://www.ietf.org/slides/slides-aic ontrolws-consent-in-crisis-the-rapid-decline-of-the-ai-data-commons-00.pdf"> <reference anchor="DECLINE" target="https://www.ietf.org/slides/slides-aic ontrolws-consent-in-crisis-the-rapid-decline-of-the-ai-data-commons-00.pdf">
<front> <front>
<title>Consent in Crisis: The Rapid Decline of the AI Data Commons</ti tle> <title>Consent in Crisis: The Rapid Decline of the AI Data Commons</ti tle>
<author initials="S." surname="Longpre" fullname="Shayne Longpre"> <author initials="S." surname="Longpre" fullname="Shayne Longpre">
skipping to change at line 243 skipping to change at line 238
</author> </author>
<author initials="A." surname="Lee" fullname="Ariel Lee"> <author initials="A." surname="Lee" fullname="Ariel Lee">
<organization/> <organization/>
</author> </author>
<author initials="C." surname="Lund" fullname="Campbell Lund"> <author initials="C." surname="Lund" fullname="Campbell Lund">
<organization/> <organization/>
</author> </author>
<date year="2025"/> <date year="2025"/>
</front> </front>
</reference> </reference>
<reference anchor="RFC9309"> <xi:include href="https://bib.ietf.org/public/rfc/bibxml/reference.RFC.930
<front> 9.xml"/>
<title>Robots Exclusion Protocol</title>
<author fullname="M. Koster" initials="M." surname="Koster"/>
<author fullname="G. Illyes" initials="G." surname="Illyes"/>
<author fullname="H. Zeller" initials="H." surname="Zeller"/>
<author fullname="L. Sassman" initials="L." surname="Sassman"/>
<date month="September" year="2022"/>
<abstract>
<t>This document specifies and extends the "Robots Exclusion Protoco
l" method originally defined by Martijn Koster in 1994 for service owners to con
trol how content served by their services may be accessed, if at all, by automat
ic clients known as crawlers. Specifically, it adds definition language for the
protocol, instructions for handling errors, and instructions for caching.</t>
</abstract>
</front>
<seriesInfo name="RFC" value="9309"/>
<seriesInfo name="DOI" value="10.17487/RFC9309"/>
</reference>
</references> </references>
<?line 245?> <?line 253?>
<section anchor="about-the-workshop"> <section anchor="about-the-workshop">
<name>About the Workshop</name> <name>About the Workshop</name>
<t>The AI-CONTROL Workshop was held on 2024-09-19 and 2024-09-20 at Wilkin son Barker Knauer in Washington DC, USA.</t> <t>The AI-CONTROL Workshop was held on 2024-09-19 and 2024-09-20 at Wilkin son Barker Knauer in Washington, D.C., USA.</t>
<t>Workshop attendees were asked to submit position papers. These papers a re published on the IAB website <xref target="PAPERS"/>, unless the submitter re quested it be withheld.</t> <t>Workshop attendees were asked to submit position papers. These papers a re published on the IAB website <xref target="PAPERS"/>, unless the submitter re quested it be withheld.</t>
<t>The workshop was conducted under the Chatham House Rule <xref target="C HATHAM-HOUSE"/>, meaning that statements cannot be attributed to individuals or organizations without explicit authorization.</t> <t>The workshop was conducted under the Chatham House Rule <xref target="C HATHAM-HOUSE"/>, meaning that statements cannot be attributed to individuals or organizations without explicit authorization.</t>
<section anchor="agenda"> <section anchor="agenda">
<name>Agenda</name> <name>Agenda</name>
<t>This section outlines the broad areas of discussion on each day.</t> <t>This section outlines the broad areas of discussion on each day.</t>
<section anchor="thursday-2024-09-19"> <section anchor="thursday-2024-09-19">
<name>Thursday 2024-09-19</name> <name>Thursday, 2024-09-19</name>
<dl> <dl>
<dt>Setting the stage</dt> <dt>Setting the stage:</dt>
<dd> <dd>
<t>An overview of the current state of AI opt-out, its impact, and existing work in this space</t> <t>An overview of the current state of AI opt-out, its impact, and existing work in this space</t>
</dd> </dd>
<dt>Lightning talks</dt> <dt>Lightning talks:</dt>
<dd> <dd>
<t>A variety of perspectives from participants</t> <t>A variety of perspectives from participants</t>
</dd> </dd>
</dl> </dl>
</section> </section>
<section anchor="friday-2024-09-20"> <section anchor="friday-2024-09-20">
<name>Friday 2024-09-20</name> <name>Friday, 2024-09-20</name>
<dl> <dl>
<dt>Opt-Out Attachment: robots.txt and beyond</dt> <dt>Opt-Out Attachment: robots.txt and beyond:</dt>
<dd> <dd>
<t>Considerations in how preferences are attached to content on th e Internet</t> <t>Considerations in how preferences are attached to content on th e Internet</t>
</dd> </dd>
<dt>Vocabulary: what opt-out means</dt> <dt>Vocabulary: what opt-out means:</dt>
<dd> <dd>
<t>What information the opt-out signal needs to convey</t> <t>What information the opt-out signal needs to convey</t>
</dd> </dd>
<dt>Discussion and wrap-up</dt> <dt>Discussion and wrap-up:</dt>
<dd> <dd>
<t>Synthesis of the workshop's topics and how future work might un fold</t> <t>Synthesis of the workshop's topics and how future work might un fold</t>
</dd> </dd>
</dl> </dl>
</section> </section>
</section> </section>
<section anchor="attendees"> <section anchor="attendees">
<name>Attendees</name> <name>Attendees</name>
<t>Attendees of the workshop are listed with their primary affiliation. Attendees from the program committee (PC) and the Internet Architecture Board (I AB) are also marked.</t> <t>Attendees of the workshop are listed with their primary affiliation. Attendees from the program committee (PC) and the Internet Architecture Board (I AB) are also marked.</t>
<ul spacing="compact"> <ul spacing="normal">
<li> <li>
<t>Jari Arkko, Ericsson</t> <t><contact fullname="Jari Arkko"/>, Ericsson</t>
</li> </li>
<li> <li>
<t>Hirochika Asai, Preferred Networks</t> <t><contact fullname="Hirochika Asai"/>, Preferred Networks</t>
</li> </li>
<li> <li>
<t>Farzaneh Badiei, Digital Medusa (PC)</t> <t><contact fullname="Farzaneh Badiei"/>, Digital Medusa (PC)</t>
</li> </li>
<li> <li>
<t>Fabrice Canel, Microsoft (PC)</t> <t><contact fullname="Fabrice Canel"/>, Microsoft (PC)</t>
</li> </li>
<li> <li>
<t>Lena Cohen, EFF</t> <t><contact fullname="Lena Cohen"/>, EFF</t>
</li> </li>
<li> <li>
<t>Alissa Cooper, Knight-Georgetown Institute (PC, IAB)</t> <t><contact fullname="Alissa Cooper"/>, Knight-Georgetown Institute (PC, IAB)</t>
</li> </li>
<li> <li>
<t>Marwan Fayed, Cloudflare</t> <t><contact fullname="Marwan Fayed"/>, Cloudflare</t>
</li> </li>
<li> <li>
<t>Christopher Flammang, Elsevier</t> <t><contact fullname="Christopher Flammang"/>, Elsevier</t>
</li> </li>
<li> <li>
<t>Carl Gahnberg</t> <t><contact fullname="Carl Gahnberg"/></t>
</li> </li>
<li> <li>
<t>Max Gendler, The News Corporation</t> <t><contact fullname="Max Gendler"/>, The News Corporation</t>
</li> </li>
<li> <li>
<t>Ted Hardie</t> <t><contact fullname="Ted Hardie"/></t>
</li> </li>
<li> <li>
<t>Dominique Hazaël-Massieux, W3C</t> <t><contact fullname="Dominique Hazaël-Massieux"/>, W3C</t>
</li> </li>
<li> <li>
<t>Gary Ilyes, Google (PC)</t> <t><contact fullname="Gary Ilyes"/>, Google (PC)</t>
</li> </li>
<li> <li>
<t>Sarah Jennings, UK Department for Science, Innovation and Technol ogy</t> <t><contact fullname="Sarah Jennings"/>, UK Department for Science, Innovation and Technology</t>
</li> </li>
<li> <li>
<t>Paul Keller, Open Future</t> <t><contact fullname="Paul Keller"/>, Open Future</t>
</li> </li>
<li> <li>
<t>Elizabeth Kendall, Meta</t> <t><contact fullname="Elizabeth Kendall"/>, Meta</t>
</li> </li>
<li> <li>
<t>Suresh Krishnan, Cisco (PC, IAB)</t> <t><contact fullname="Suresh Krishnan"/>, Cisco (PC, IAB)</t>
</li> </li>
<li> <li>
<t>Mirja Kühlewind, Ericsson (PC, IAB)</t> <t><contact fullname="Mirja Kühlewind"/>, Ericsson (PC, IAB)</t>
</li> </li>
<li> <li>
<t>Greg Leppert, Berkman Klein Center</t> <t><contact fullname="Greg Leppert"/>, Berkman Klein Center</t>
</li> </li>
<li> <li>
<t>Greg Lindahl, Common Crawl Foundation</t> <t><contact fullname="Greg Lindahl"/>, Common Crawl Foundation</t>
</li> </li>
<li> <li>
<t>Mike Linksvayer, GitHub</t> <t><contact fullname="Mike Linksvayer"/>, GitHub</t>
</li> </li>
<li> <li>
<t>Fred von Lohmann, OpenAI</t> <t><contact fullname="Fred von Lohmann"/>, OpenAI</t>
</li> </li>
<li> <li>
<t>Shayne Longpre, Data Provenance Initiative</t> <t><contact fullname="Shayne Longpre"/>, Data Provenance Initiative< /t>
</li> </li>
<li> <li>
<t>Don Marti, Raptive</t> <t><contact fullname="Don Marti"/>, Raptive</t>
</li> </li>
<li> <li>
<t>Sarah McKenna, Alliance for Responsible Data Collection; Sequentu m</t> <t><contact fullname="Sarah McKenna"/>, Alliance for Responsible Dat a Collection; Sequentum</t>
</li> </li>
<li> <li>
<t>Eric Null, Center for Democracy and Technology</t> <t><contact fullname="Eric Null"/>, Center for Democracy and Technol ogy</t>
</li> </li>
<li> <li>
<t>Chris Needham, BBC</t> <t><contact fullname="Chris Needham"/>, BBC</t>
</li> </li>
<li> <li>
<t>Mark Nottingham, Cloudflare (PC)</t> <t><contact fullname="Mark Nottingham"/>, Cloudflare (PC)</t>
</li> </li>
<li> <li>
<t>Paul Ohm, Georgetown Law (PC)</t> <t><contact fullname="Paul Ohm"/>, Georgetown Law (PC)</t>
</li> </li>
<li> <li>
<t>Braxton Perkins, NBC Universal</t> <t><contact fullname="Braxton Perkins"/>, NBC Universal</t>
</li> </li>
<li> <li>
<t>Chris Petrillo, Wikimedia</t> <t><contact fullname="Chris Petrillo"/>, Wikimedia</t>
</li> </li>
<li> <li>
<t>Sebastian Posth, Liccium</t> <t><contact fullname="Sebastian Posth"/>, Liccium</t>
</li> </li>
<li> <li>
<t>Michael Prorock</t> <t><contact fullname="Michael Prorock"/></t>
</li> </li>
<li> <li>
<t>Matt Rogerson, Financial Times</t> <t><contact fullname="Matt Rogerson"/>, Financial Times</t>
</li> </li>
<li> <li>
<t>Peter Santhanam, IBM</t> <t><contact fullname="Peter Santhanam"/>, IBM</t>
</li> </li>
<li> <li>
<t>Jeffrey Sedlik, IPTC/PLUS</t> <t><contact fullname="Jeffrey Sedlik"/>, IPTC/PLUS</t>
</li> </li>
<li> <li>
<t>Rony Shalit, Alliance For Responsible Data Collection; Bright Dat a</t> <t><contact fullname="Rony Shalit"/>, Alliance For Responsible Data Collection; Bright Data</t>
</li> </li>
<li> <li>
<t>Ian Sohl, OpenAI</t> <t><contact fullname="Ian Sohl"/>, OpenAI</t>
</li> </li>
<li> <li>
<t>Martin Thomson, Mozilla</t> <t><contact fullname="Martin Thomson"/>, Mozilla</t>
</li> </li>
<li> <li>
<t>Thom Vaughan, Common Crawl Foundation (PC)</t> <t><contact fullname="Thom Vaughan"/>, Common Crawl Foundation (PC)< /t>
</li> </li>
<li> <li>
<t>Kat Walsh, Creative Commons</t> <t><contact fullname="Kat Walsh"/>, Creative Commons</t>
</li> </li>
<li> <li>
<t>James Whymark, Meta</t> <t><contact fullname="James Whymark"/>, Meta</t>
</li> </li>
</ul> </ul>
<t>The following participants requested that their identity and/or affil iation not be revealed:</t> <t>The following participants requested that their identity and/or affil iation not be revealed:</t>
<ul spacing="compact"> <ul spacing="normal">
<li> <li>
<t>A government official</t> <t>A government official</t>
</li> </li>
</ul> </ul>
</section> </section>
</section> </section>
<section numbered="false" anchor="iab-members-at-the-time-of-approval"> <section numbered="false" anchor="iab-members-at-the-time-of-approval">
<name>IAB Members at the Time of Approval</name> <name>IAB Members at the Time of Approval</name>
<t>Internet Architecture Board members at the time this document was appro ved for publication were:</t> <t>Internet Architecture Board members at the time this document was appro ved for publication were:</t>
<ul spacing="compact"> <ul spacing="normal">
<li> <li>
<t>Matthew Bocci</t> <t><contact fullname="Matthew Bocci"/></t>
</li> </li>
<li> <li>
<t>Roman Danyliw</t> <t><contact fullname="Roman Danyliw"/></t>
</li> </li>
<li> <li>
<t>Dhruv Dhody</t> <t><contact fullname="Dhruv Dhody"/></t>
</li> </li>
<li> <li>
<t>Jana Iyengar</t> <t><contact fullname="Jana Iyengar"/></t>
</li> </li>
<li> <li>
<t>Cullen Jennings</t> <t><contact fullname="Cullen Jennings"/></t>
</li> </li>
<li> <li>
<t>Suresh Krishnan</t> <t><contact fullname="Suresh Krishnan"/></t>
</li> </li>
<li> <li>
<t>Mirja Kühlewind</t> <t><contact fullname="Mirja Kühlewind"/></t>
</li> </li>
<li> <li>
<t>Warren Kumari</t> <t><contact fullname="Warren Kumari"/></t>
</li> </li>
<li> <li>
<t>Jason Livingood</t> <t><contact fullname="Jason Livingood"/></t>
</li> </li>
<li> <li>
<t>Mark Nottingham</t> <t><contact fullname="Mark Nottingham"/></t>
</li> </li>
<li> <li>
<t>Tommy Pauly</t> <t><contact fullname="Tommy Pauly"/></t>
</li> </li>
<li> <li>
<t>Alvaro Retana</t> <t><contact fullname="Alvaro Retana"/></t>
</li> </li>
<li> <li>
<t>Qin Wu</t> <t><contact fullname="Qin Wu"/></t>
</li> </li>
</ul> </ul>
</section> </section>
<section numbered="false" anchor="acknowledgements"> <section numbered="false" anchor="acknowledgements">
<name>Acknowledgements</name> <name>Acknowledgements</name>
<t>The Program Committee and the IAB would like to thank Wilkinson Barker Knauer for their generosity in hosting the workshop.</t> <t>The program committee and the IAB would like to thank Wilkinson Barker Knauer for their generosity in hosting the workshop.</t>
<t>We also thank our scribes for capturing notes that assisted in the prod uction of this report:</t> <t>We also thank our scribes for capturing notes that assisted in the prod uction of this report:</t>
<ul spacing="compact"> <ul spacing="normal">
<li> <li>
<t>Zander Arnao</t> <t><contact fullname="Zander Arnao"/></t>
</li> </li>
<li> <li>
<t>Andrea Dean</t> <t><contact fullname="Andrea Dean"/></t>
</li> </li>
<li> <li>
<t>Patrick Yurky</t> <t><contact fullname="Patrick Yurky"/></t>
</li> </li>
</ul> </ul>
</section> </section>
</back> </back>
<!-- ##markdown-source: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</rfc> </rfc>
 End of changes. 106 change blocks. 
397 lines changed or deleted 192 lines changed or added

This html diff was produced by rfcdiff 1.48.