Lab Weekly — 05/23/2025 Google goes AI Mode; Plus, the latest news about agentic web, AI audio product summaries, and more must-know stories and stats

Want this newsletter in your inbox every week? Sign up for our mailing list here

Google I/O 2025 recap: AI Mode, Activated

At I/O 2025, Google unveiled its boldest vision yet for an AI-first future; Here’s what brands need to know

In case you missed it… What Brands Can Learn from the Labubu Fad

Decoding the Labubu phenomenon, the “Little Treat” Culture, and the gamble-fication of everything

How “A Minecraft Movie” Hit Box Office Gold

Decoding how the latest video game adaptation broke out and became a huge hit by riding on memes, deeply engaging its fan base, while also appealing to broader audiences

Airbnb’s Rebranding & What It Says About the Service Economy

The issue with scaling local services and experiences, and how AI might play a part in Airbnb’s future

Episode 170: AI Mode, Agentic Web, & more from Google I/O 2025

On the latest Floor 9 episode, four members of the Lab team — Tom, Ryan, Richard, and Elisa — assemble to discuss in-depth about everything brand marketers that need to know about the AI-centric announcements coming out of Google’s annual developer event.

From Google’s ambition of using AI Mode to transform its search products, to the “agentic web” that the AI stakeholders are aiming towards, the team places the various key announcements within the larger industry context and discuss their implications for consumers and brands alike. Listen now to learn more about how you brand should respond to the latest developments in consumer-facing AI.

Microsoft Launches NLWeb to Bring More Chatbots to Web Pages [TechCrunch]

As part of its goal to facilitate the future of an “agentic web” where consumers mostly interact via AI chatbots, Microsoft is launching NLWeb, an open-source protocol designed to simplify the integration of AI-powered conversational interfaces across the web. With just a few lines of code, websites can now embed a chatbot-like experience — powered by the AI model of their choice and fed with their own data. A travel retailer, for instance, might use NLWeb to help customers find outfits for different destinations, while a food blog could use it to recommend wine pairings or complementary recipes.

Microsoft likens NLWeb’s role in the emerging “agentic web” to HTML’s role in the graphical web, allowing users to interact directly with content using natural language. Of course, not every brand needs to incorporate a full-scale AI chatbot into their brand experience. Instead, look for low-risk, high-utility entry points — for example, a conversational layer that helps users find support articles, or spark new inspirations based on pop culture references. These narrow use cases allow for controlled testing, better user feedback, and iteration without overextending resources.

Related: Microsoft is opening its on-device AI models up to web apps in Edge [The Verge]; Microsoft’s Edit on Windows is a new command-line text editor [The Verge]; Google’s AI agents will bring you the web now [TechCrunch]

Amazon Rolls Out Short-Form AI-Powered Audio Product Summaries [TechCrunch]

Amazon is bringing short-form AI-powered audio product summaries to select product pages, aiming to help Amazon customers save time while shopping and get important information through a conversational, discussion-style format. The new AI product summaries can be accessed by tapping the “Hear the highlights” button in the Amazon app.

This move reflects a broader trend toward voice-first and multimodal shopping experiences, where convenience, accessibility, and personalization are prioritized. For Amazon, it’s also a strategic step in making product discovery more engaging — particularly for on-the-go users, visually impaired customers, or those overwhelmed by lengthy product descriptions and reviews. I do wonder what kind of role that Alexa could play in this type of voice-driven shopping experiences that Amazon is clearly interested in testing out.

Related: Google launches NotebookLM for iOS and Android [9to5Google]; Weeks after Amazon’s Alexa+ AI launch, a mystery: where are the users? [Reuters]

YouTube Will Air Its First Exclusive NFL Game From Brazil [MSN]

YouTube will air its first exclusive NFL game — a Week 1 matchup between the Kansas City Chiefs and Los Angeles Chargers — from São Paulo, Brazil on September 5. The game will stream globally for free, marking another step in the NFL’s strategy to reach younger, international audiences through digital platforms. This development highlights two converging trends: the globalization of U.S. professional sports and the accelerating shift toward streaming. By staging a marquee game in Brazil and broadcasting it on a globally accessible platform like YouTube, the NFL is actively cultivating international fan bases and acknowledging that future growth lies beyond U.S. borders.

At the same time, the fragmentation of sports content across streaming platforms (YouTube, Amazon, Netflix, ESPN+, etc.) poses challenges for viewers who now must navigate multiple services to follow their favorite teams. This shift, while opening new monetization and engagement opportunities, could lead to subscription fatigue and dilute the unified experience once offered by traditional broadcast networks.

Related: YouTube launches weekly top podcast list to rival Spotify and Apple [CNBC]; Sesame Street’s next season will stream on Netflix [The Verge]

Situational Awareness: OpenAI’s Codex brings coding support directly into ChatGPT [ZDNET]

Coding has now proven to be the most low-hanging fruit for LLM models. Vibe-coding is absolutely becoming a thing now with the advance of AI-powered coding tools. And between GItHub, Google, and now OpenAI, we do wonder whether this democratization of coding and software development may lead to a transformation in how tech startups are founded and operated, potentially accelerating innovation and reducing development costs. Furthermore, as AI coding tools become more prevalent, we might see a more evenly distributed landscape in 3D content creation and metaverse development.

Jony Ive to lead OpenAI’s design work [TechCrunch]

There’s some irony/poetic justice to think that the designer-in-chief who’s responsible to craft the first generations of iPhones and usher in the mobile era might also be the one to unlock a new generation of AI-first devices. The latest reports say that OpenAI’s next “big project” will not be a wearable device; Instead, it will be a compact, screenless device, fully aware of its user’s surroundings. Sounds like the kind of futuristic challenge that they will need Jony Ive for!

Strava is buying up athletic training apps, Runna and Breakaway [TechCrunch]

Fitness-oriented social network Strava has made two acquisitions over the last month: cycling app The Breakaway and Runna, a leading running app. These strategic acquisitions aim to enhance Strava by integrating specialized training tools that the fitness communities already use and care about. Once integrated, these tools can provide Strava users with personalized coaching and training plans, thereby increasing user engagement and retention. Between this and Letterboxd launching an “TVOD” movie rental store, we could see other interest-based networks follow suit and take similar routes to broaden their offerings.

Chicago Sun-Times prints summer reading list full of fake books [ARSTechnica]

The Chicago-based publisher published a summer reading list over the weekend that included at least 10 fake book titles attributed to real, well-known authors. This faux-pas points to the fact that AI hallucination is still very much an issue and human oversight is still much needed to avoid embarrassing mishaps like this. While AI tools can be useful for drafting and ideation, brands should be cautious when it comes to its execution and deployment.

  • Despite the faltering metaverse for consumers, the World Economic Forum projects the industrial metaverse is expected to be worth $100 billion globally by 2030, Wired reports. Analysts predicts that the biggest impact from the industrial metaverse will be embodied or physical AI — in other words, robots.
  • Researchers at found that AI chatbot can be quite persuasive if they know some personal information about their target. When provided with minimal demographic information on their opponents, AI chatbots — known as large language models (LLMs) — were able to adapt their arguments and be more persuasive than humans in online debates 64 percent of the time, according to a study published in Nature Human Behavior, The Washington Post reports.

If you find our insights valuable and would like to have a deeper conversation on technology and media innovations, or need to sound smarter in a client meeting or a pitch, please feel free to reach out to Chelsea Freitas, our VP of Strategy, at chelsea@ipglab.com.

If you liked this edition of Lab Weekly by IPG Media Lab, why not share it?

Lab Weekly — 05/23/2025 was originally published in IPG Media Lab on Medium, where people are continuing the conversation by highlighting and responding to this story.