Generative AI 2025 🆕

[nrllh]

trafficstars

Generative AI 2025

If you're interested in contributing to the Generative AI chapter of the 2025 Web Almanac, please reply to this issue and indicate which role or roles best fit your interest and availability: author, reviewer, analyst, and/or editor. You can find more details about this year’s Call for Contributions here.

💬 Please make sure to join #web-almanac on Slack for team coordination.

Content team

Lead	Authors	Reviewers	Analysts	Editors	Coordinator
@christianliebel	@christianliebel, @Yash-Vekaria, @blogsmithteam @jongoodey @mikaelaraujo	@Yash-Vekaria @VaheSODP @UmarIqbal	@VaheSODP @isabellajantti @JonathanPLev	@blogsmithteam	@nrllh

Expand for more information about each role 👀

• The content team lead is the chapter owner and responsible for setting the scope of the chapter and managing contributors' day-to-day progress.
• Authors are subject matter experts and lead the content direction for each chapter. Chapters typically have one or two authors. Authors are responsible for planning the outline of the chapter, analyzing stats and trends, and writing the annual report.
• Reviewers are also subject matter experts and assist authors with technical reviews during the planning, analyzing, and writing phases.
• Analysts are responsible for researching the stats and trends used throughout the Almanac. Analysts work closely with authors and reviewers during the planning phase to give direction on the types of stats that are possible from the dataset, and during the analyzing/writing phases to ensure that the stats are used correctly.
• Editors are technical writers who have a penchant for both technical and non-technical content correctness. Editors have a mastery of the English language and work closely with authors to help wordsmith content and ensure that everything fits together as a cohesive unit.
• The section coordinator is the overall owner for all chapters within a section like "User Experience" or "Page Content" and helps to keep each chapter on schedule.

Note: The time commitment for each role varies by the chapter's scope and complexity as well as the number of contributors.

For an overview of how the roles work together at each phase of the project, see the Chapter Lifecycle doc.

Milestone checklist

0. Form the content team

• [x] 📆 May 18 Complete program and content committee - 🔑 Organizing committee
  • The content team has at least one author, reviewer, and analyst.

1. Plan content

• [x] 📆 June 1 First meeting to outline the chapter contents - 🔑 Content team
  • The content team has completed the chapter outline.

2. Gather data

• [ ] 📆 July 1 Custom metrics completed - 🔑 Analysts
  • Analysts have added all necessary custom metrics and drafted a PR (example) to track query progress.
• [ ] 📆 July 1 HTTP Archive Crawl - 🔑 HA Team
  • HTTP Archive runs the June crawl.

3. Validate results

• [ ] 📆 September 1 Query Metrics & Save Results - 🔑 Analysts
  • Analysts have queried all metrics and saved the output.

4. Draft content

• [ ] 📆 October 1 First Draft of Chapter - 🔑 Authors
  • Authors has written the chapter.
• [ ] 📆 October 20 Review & Edit Chapter - 🔑 Reviewers & Editors
  • Reviewers and Editors has processed the the chapter.

5. Publication

• [ ] 📆 November 15 Chapter Publication (Markdown & PR) - 🔑 Authors
  • Authors has converted the chapter to markdown and drafted a PR.
• [ ] 📆 December 1 Launch of 2025 Web Almanac 🚀 - 🔑 Organizing committee

6. Live Stream

• [ ] 📆 December 15 Live Stream - 🔑 Content Team

Chapter resources

Refer to these 2025 generative-ai resources throughout the content creation process:

• 📄 Planning doc for outlining and drafting content
• 📊 Results sheet for saving the results of queries
• 📝 Markdown file for publishing content and managing public metadata
• 🔍 SQL files for committing the queries used during analysis
• 💬 #web-almanac-generative-ai on Slack for team coordination

[Yash-Vekaria]

This is super exciting! Happy to join this chapter as author and/or reviewer.

[max-ostapenko]

I'll participate as an analyst.

[VaheSODP]

interested as an analyst or reviewe

[blogsmithteam]

I'm interested in being an author or can support as an editor :)

[nrllh]

Cool folks, we’re shaping up to be a great team!

[ghost]

Keen to jump in as an analyst if there's still a need for one :)

[jongoodey]

I hope I'm doing this right, but I would like to join this chapter as author.

[nrllh]

This chapter is shaping up really very well - I’m really excited about this chapter!

I’ve assigned @isabellajantti as analyst, @jongoodey as author, and @UmarIqbal as reviewer.

Thank you all!

[christianliebel]

Wow, it’s great to see so many people interested in the topic and willing to contribute.

@Yash-Vekaria @blogsmithteam @jongoodey @VaheSODP @UmarIqbal @max-ostapenko @isabellajantti, @nrllh: I can offer a Zoom meeting on May 28, 2025, at 4:00 PM UTC (9:00 AM PDT, 5:00 PM BST, 6:00 PM CEST) as a kickoff for our chapter. Unfortunately, this is the only date I can provide before the June 1 deadline that accommodates the APAC/EMEA regions, where most contributors are located according to GitHub bios.

In the meantime, here are my ideas about potential data points/metrics:

• Use of OpenAI/… cloud services (https://api.openai.com/v1/*, …)
• … maybe check if people leak sk-* keys (OpenAI and other services)?
• Use of WebNN (e.g., MLGraphBuilder interface)
• Use of WebLLM (e.g., ChatModule class)
• Use of Transformers.js , ONNX Runtime Web or TensorFlow.js (+ versions?)
• Use of Built-in AI APIs (Prompt API, Writing Assistance APIs, Translator and Language Detector APIs)
• Models requested/used
• Local: Size of models downloaded

Feel free to add any other ideas.

[JonathanPLev]

I would love to jump onto this chapter as an analyst if still possible.

[nrllh]

Can you please join our Slack channel #web-almanac-generative-ai for this chapter?

[mikaelaraujo]

@nrllh I want to contribute as an author, if possible.

[nrllh]

This is a kindly reminder for the Zoom meeting on May 28, 2025, at 4:00 PM UTC (9:00 AM PDT, 5:00 PM BST, 6:00 PM CEST) as a kickoff for this chapter.

cc @christianliebel @Yash-Vekaria @blogsmithteam @jongoodey @VaheSODP @UmarIqbal @max-ostapenko @isabellajantti @JonathanPLev @mikaelaraujo

@christianliebel you might want to share the Zoom link or a calendar invite (.ics) here or in Slack.

[christianliebel]

I will share the Zoom link in our Slack channel.

[christianliebel]

@Yash-Vekaria @blogsmithteam @jongoodey @VaheSODP @UmarIqbal @max-ostapenko @isabellajantti @JonathanPLev @mikaelaraujo Reminder for our meeting in ~50 mins.

[blogsmithteam]

Traveling but aiming to join!

[blogsmithteam]

@christianliebel can you drop the zoom link here or email it to [email protected]? I think I'm going to need help to access Slack because I don't have one of these email suffixes:

[mikaelaraujo]

@christianliebel can you drop the zoom link here or email it to [email protected]? I think I'm going to need help to access Slack because I don't have one of these email suffixes:

@blogsmithteam, I sent an invitation to your email. Please, check.

[max-ostapenko]

I took a lead role in the Privacy chapter additionally, unfortunately will not have enough time to support with analysis here. Please unassign me. The team has great analytical support, so I'm looking forward to reading an insightful chapter.

[christianliebel]

I took a lead role in the Privacy chapter additionally, unfortunately will not have enough time to support with analysis here. Please unassign me.

Oh, okay. I’ve unassigned you. Thanks for your help so far.

[christianliebel]

Dear analysts (@VaheSODP, @isabellajantti, @JonathanPLev), now is your time. Could you please assign yourselves to the metrics in our planning doc?

Dear authors (@Yash-Vekaria, @blogsmithteam, @jongoodey, @mikaelaraujo), if you wish to start early, for example, by writing the introduction or sections unrelated to the result data, feel free to place your drafts in the planning doc.

[clarkio]

Hi everyone, I've typically contributed to the security chapter each year but am interested in this chapter this year as well. I could help as en Editor and/or Reviewer here. Let me know if you could use the help and if not no worries 👍

[christianliebel]

@clarkio Sounds great! If you have the capacity, you would be welcome to join both as an editor and a reviewer. If you only have the capacity for one role, I would suggest joining as an editor, as we only have one for now.

[christianliebel]

Gentle reminder for our analysts (@VaheSODP, @isabellajantti, @JonathanPLev): Currently, only one custom metric is assigned to someone.

@clarkio, what is your preference? Would you like to join both as an editor and a reviewer?

[KenjiBaheux]

Hi everyone,

I'm very much looking forward to this new chapter! I'd be happy to contribute as an editor or reviewer if you need extra help.

Question on Chapter Scope

I have a quick question regarding the chapter's scope. The title focuses on "Generative AI," but I noticed the plan mentions technologies like the Translator and Language Detector APIs, which don't necessarily use generative models (e.g., in Chrome's current implementation).

Is the intent to focus strictly on generative use cases, or could this be a broader "AI/ML on the Web" chapter? Personally, I think a broader scope would be highly valuable, but I wanted to check if that topic is already planned for another section.

Suggestions for Metrics and Analysis

If it's not too late to contribute ideas, here are a few thoughts on what could be valuable to measure. I recognize that some of these may be challenging or impossible to capture with the current options, but wanted to share them in case they are feasible or could inform a discussion for better tooling / data with the next iteration:

• Categorization of AI/ML Usage: It would be insightful to break down usage by:

  • Use case: (e.g., text generation, image effects, image enhancement, recommendation, translation).
  • Modality: (e.g., text, image, audio, multimodal).
  • Execution environment: Client-side, server-side, or hybrid.
  • Site vertical: (e.g., e-commerce, media, SaaS).
  • Site scale: Usage patterns on "head" (top 10k sites) vs. "tail" (long-tail).

• Underlying Web Platform Adoption:

  • Beyond tracking the new WebNN API, it would be powerful to measure WebGPU and Wasm usage, and if possible, attempt to differentiate AI/ML workloads from other uses (perhaps by correlating with the presence of ML libraries).

• Framework and Model Details:

  • I strongly support the suggestion to track framework versions. It would also be useful to know how they are served (e.g., from a canonical CDN vs. self-hosted).
  • For AI/ML frameworks like transformers.js, could we track how often developers specify a model versus relying on the defaults (e.g., by tracking pipeline usage without an explicit model argument vs. overriding the default)?
  • For the models themselves, could we identify variations like version or quantization level (e.g., float32 vs. float16) and their delivery method? For the former, popularity heatmaps of models would be really neat (e.g. model families, convergence on a handful of versions and/or quantizations versus a heavily scattered landscape).

Looking forward to seeing this chapter develop!

[webmaxru]

Hello folks! I'm abit late but I'd love to join as a reviewer if there is more slots left. I do lots of experimentation and public speaking around WebNN, and I have experience of being HTTP Archive chapters reviewer several times.

[christianliebel]

@webmaxru Sure, I have added you as a reviewer.

[christianliebel]

@KenjiBaheux Hi Kenji, thanks for your insights, and please excuse my late reply.

Question on Chapter Scope: This chapter is indeed not strictly about GenAI, but also about other topics regarding AI/ML on the web. I'm unsure if changing the chapter's title is still possible (paging @nrllh), but it would make sense given our ideas for the following years.

Suggestions for Metrics and Analysis: Unfortunately, most analysts had to withdraw from the chapter for various reasons. We currently only have metrics for the Built-in AI APIs, llm.txt, and the Chrome Platform Status data for WebGPU and Wasm. We can definitely take your suggestions into account for next year.

You're welcome to join the chapter as an editor or reviewer. Please tell me which role you prefer.

[christianliebel]

Dear authors (@Yash-Vekaria @blogsmithteam @jongoodey @mikaelaraujo), would you be ready to start writing?

[mikaelaraujo]

@christianliebel Sure.