Peggy K's Creator Weekly: Google I/O, AI Everywhere, Simplified Pinterest Pinning

Peggy K's Weekly Newsletter Header

Happy Mother’s Day weekend! 🌹

This week was Google I/O, Google’s big developer conference, which was a showcase of how Google is integrating AI across its products and services. Most of the new features won’t be available for a while, but there are a few you can try today (if available in your country). Read on for details!

Plus there’s more non-Google news and updates for video creators, writers and more.

Note: Last week’s newsletter may have ended up in your spam folder marked as possibly malicious. I suspect it was the discussion and links for passkeys that may have triggered that. Apologies!

Creator Weekly Live

Join me live on Sunday for a recap of this week’s top updates! It’s live at 10:30AM Pacific time (5:30 UTC). Set a reminder or join on YouTube here.

Thanks to everyone who has joined me live the past few weeks as I learn more about going live. I appreciate the folks in live chat!

Google’s AI Everywhere

Google has been criticized by being slow to launch new AI features while Microsoft and other companies have quickly launched chatbots and other AI-driven tools. The keynote address at this week’s Google I/O seems to have put most of those criticisms to rest.

The Verge compiled a video of every time Google CEO Sundar Pichai said “AI” during the Google I/O keynote.

Google has made clear that they are taking a considered responsible approach. For example, large language models and chatbots can sometimes give responses almost human-like, with emotional or opinionated-seeming responses. Google says they have trained theirs “to refrain from reflecting a persona”, so it will provide “objective neutral responses” and be less conversational. And it will not respond or add a disclaimer in response to questions on sensitive or “your money or your life” (YMYL) topics. You can read more here (pdf file).

Many of these AI features are built on Google’s next generation language model PaLM2, which is trained on multilingual text, scientific papers, and source code.

I’ll note that there were a number of other non-AI announcements at II/O that I’m not going to cover, including a new folding phone and a tablet, updates to WearOS (for smart watches), Google Home (for controlling smart devices in your home), Android Auto, immersive view in Maps and more.

For an overview see:

And now for the AI integrations. I’m including information about when and how you can try out the new features. Note that some may only be available in limited countries and languages.

Google is integrating generative AI into Search, to “take more of the work out of searching.” Enter a question into search and you can get an AI-written overview of the information, including a few links. You’ll then be able to ask follow-up questions and see suggested next steps. It remains to be seen whether this will hurt the websites that provided the information the answer is based on. (And yes, ads will be inserted in “dedicated ad slots” on the page).
How can you try it? Sign up for the new Search Labs to test out new features over the coming weeks. It’s available in the US, English only.

When you use Google Search to shop for something, you can get up-to-date product description and reviews with generative AI trained on information from Google’s constantly updated “Shopping Graph”.
How can you try it? Sign up for the new Search Labs to test out new features over the coming weeks. It’s available in the US, English only.

AI-generated images (at least those created by Google) will have markup so that they are labeled as AI-generated in the Google Image Search results. Creators and publishers (like Shutterstock and Midjourney) will be able to add that markup as well.

Google Bard is Google’s chatbot. It has already been updated a number of times since its launch last March. It is now available in English, Japanese and Korean in more than 180 countries (but not Europe outside the UK), and content generated by Bard can be exported to Gmail and Docs. Features coming soon include the ability to include images in a prompt (which is understood using Google Lens technology), images in responses, better source citations, and the ability to export code to Replit.
How can you try it? Sign up for Google Bard.

In the coming months there will be integration of Bard with Adobe Firefly to generate images from text prompts which you can edit in (free with account) Adobe Express. This is notable because Adobe trained Firefly on licensed and public domain images, rather than just scraping the internet.
How can you try it? Sign up for Google Bard and try the Adobe Firefly beta.

Duet AI is the name for Google’s AI-based productivity tools for Google Workspace. The idea is that you are working in harmony with the AI.

  • Create a text draft in Gmail or Docs with a short prompt. In Docs it will include smart chips for information like your company name, locations or status to make them easy to customize across the document.
    How can you try it? This was originally announced last March for desktop, and it’s soon going to be available in Gmail on mobile. Sign up for Workspace Labs to try it. (It should be available for consumer accounts, as well as Workspace accounts.)

  • In Google Slides you will be able to use a new “Help me Visualize” tool to use a text prompt to generate images.
    How can you try it? This should eventually be available in Workspace Labs.

  • In Google Sheets AI can classify and label data in a Sheet, and a new “Help me organize” tool will “create a custom plan for tasks, projects, or any activity that you want to track or manage”.
    How can you try it? This should eventually be available in Workspace Labs.

  • In Google Meet you will be able to generate unique custom backgrounds for your video meetings.
    How can you try it? This should eventually be available in Workspace Labs.

Project Tailwind is an “AI-first notebook, powered by your notes and sources.”
How can you try it? Sign up for the waiting list. US only.

There was a “sneak peek” at Google Photos Magic Editor, which uses generative AI to let you reposition the subject of a photo. The AI creates new content to fill in the gaps.
How can you try it? “Select” Pixel phones will have access later this year.

There was a demo of a Universal Translator that can take a video, create new translated audio, then matches the speaker’s lips so it looks more natural than dubbed.
How can you try it? You can’t. This is apparently an experiment, and was used as an example of an AI tool that has great potential, but could be high risk in the hands of bad actors.

Messages by Google -- the Android text messaging and RCS app -- is getting Magic Compose. It’s a souped-up version of smart replies, that let you choose the tone, from business-like to fun.
How can you try it? It will be available “this summer” in beta. Sign up to be a Messages beta tester.

Use Generative AI to create wallpaper for your Android phone. Customize text-to-image prompts to create a new wallpaper for your phone. Also new: emoji wallpaper and “cinematic” wallpaper that gives your own photo a 3D look.
How can you try it? These are coming to Pixel devices next month. No announcement when they may be more widely available.

MusicLM turns text descriptions into music (like “soulful jazz for a dinner party” or “ambient, soft sounding music I can study to”).
How can you try it? This is available in Google’s AI Test Kitchen.

Code Tips uses AI large language models to write code faster and smarter. You can ask questions about programming languages (C, C++, Go, Java, JavaScript, Kotlin, Python, TypeScript), tools (Docker, Git, shells), and algorithms.
How can you try it? Sign up for the new Search Labs to test out new features over the coming weeks. It’s available in the US, English only.

Project Starline is Google’s life-size 3D video calling platform. The original prototype used hardware that took up a small room. AI has aided development of a new prototype that isn’t much bigger than a large computer monitor.
How can you try it? Google is currently testing Starline internally and with some businesses like Salesforce. No word on when it will be more widely available.

Project Gameface is an open-source hands-free, AI-powered gaming mouse. The mouse is controlled by facial expressions and gestures.
How can you try it? If you are interested in contributing to the project or trying the technology, check out the code repository on GitHub.

Finally, Google offers enterprise access to their Large Language Model and generative AI through the Google Cloud Vertex AI. Companies working with Google Cloud include Box, Canva, Dialpad, Jasper, Salesforce and UKG. Learn more about what they are using it for.

More Recent AI News

Geoffrey Hinton, a pioneer in neural networks (and so-called “godfather of AI”), left his job at Google so he could talk more freely about the dangers of AI. A major concern is that an AI arms race between Microsoft, Google and other companies could cause a flood of false AI-generated content. He apparently thinks there should be more controls on AI development in general as AI eventually could be smarter than humans.

AI researcher (and former Google employee) Meredith Whittaker is disappointed by Hinton’s talk tour, as he apparently downplayed internal concerns about AI while at Google. Her concern is that AI can (and does) disproportionately impact already marginalized people, while there is no near-term threat that current AI models have a “spark of consciousness”.

Casey Newton @ Platformer explains why he’s having trouble covering AI. Cover the new fun features? Or focus on the potential risk? There needs to be a way to include both.

Microsoft is integrating AI more deeply with search in Bing and Edge. There will be more images and videos in answers, in addition to text, and are adding multi-session options for productivity (instead of single sessions). Also coming soon: export and share conversations and improved summarization of long documents. Microsoft is also opening up Bing chat to developers.

Microsoft’s Image Creator in Bing chat now supports more languages.

Video Updates

This week at YouTube: Expiring Community posts on Android and iOS, new filter in the Creator Music Library for license duration. Learn more from Creator Insider.

Apple is bringing its Final Cut Pro video editor and Logic Pro music editor to the iPad. This looks like a great update for creators who want professional-grade software to edit on the go.

App developers and businesses can use the new Google Maps Aerial View to incorporate 3D cinematic video into applications.

Facebook is expanding Ads on Reels to include more creators. Creator earnings are based on the number of plays, rather than offering a cut of the ad revenue.

TikTok launched a #NewMedia Search Hub, “a dedicated space for artists around the world to showcase their new songs and for fans to discover their next favorite track.”

As part of Pinterest’s new streamlined Pin creation process, all video Pins will now have enhanced metrics, including views and watch time.

StreamYard now lets you customize your branding for live On-Air webinars (requires paid Professional plan). And there is now automatic starring of Super Chats and new member chats in live streams.

Spotify for Artists explains how to “make the most” of the Now Playing view, by adding visuals, tagging merch, and making sure your profile is up to date.

Web Publishers

Google is updating their search ranking algorithms “with a greater focus on content with unique expertise and experience”. Google doesn’t share how exactly their system “deeply understands content created from a personal or expert point of view”, but it’s meant to surface harder to find content, such as “a comment in a forum thread, a post on a little-known blog, or an article with unique expertise on a topic.” It sounds like it could be good for the little guys. We’ll see. (Also see Google’s new Perspectives filter, noted under Social Media updates).

Google image search will show “About this Image”, with information about where it first appeared, where else it is found online, and when it was first indexed by Google.

Social Media

Google is launching a new Perspectives filter for the Search results. This will appear when you are searching for something that “might benefit from the experiences of others”. This will pull posts and information from social media, forums, message boards and Q & A sites. It will include the name, profile photo, and other details about the creator of the content.

Pinterest is simplifying their Pin creation flow and making features available to all creators. Now all Pins can have links, music, text overlay, and stickers. Monetization is more widely available, with the paid partnership tool and tagging with affiliate links are now available to anyone. And you’ll be able to edit Pins after posting.

Twitter’s new “inactive account” policy will close inactive accounts. What does “inactive” mean exactly? Twitter suggests you sign in to your account at least every 30 days (!)

Twitter will be launching a new verification option for small businesses that is less than the current $1000 per month.

Is the Meta Metaverse headed towards the graveyard? Ed Zitron @ Business Insider thinks so.

LinkedIn users may see a follower drop after shutting down its job search platform in China, and no longer including Restricted or Hibernated accounts in the total number of followers and connections.

Productivity

Google is testing an Add to Sheets tool that lets you add links from the Google Search results into a Google Sheets spreadsheet. Sheets can easily be shared with friends, family and colleagues. Sign up for the new Search Labs to give that (and other search experiments) a try. It’s available in the US only.

The Photopea image editor has a new Magic Remove tool, that lets you remove unwanted people or objects from your photos.

Google Drive has a new spam folder. Google will detect if a file with spam or abusive content is shared with you and move it to the spam folder. Spam will be deleted after 30 days.

More Reading

I was sad to read that Heather Armstrong, AKA Dooce, died by suicide. Obituaries are referring to her as a “mommy blogger”, and she was that, but it doesn’t really capture how influential she was during those golden years of blogging in the early aughts. “Dooced” became part of the online lexicon when she was apparently fired from her original job for writing about her work and colleagues. And I think much of her popularity was because she wrote frankly about her struggles with motherhood and depression. Her final post, in April, was about being sober for 18 months and appreciating her now-18-year-old daughter.

This summer Google is launching unknown tracker alerts on Android devices. This will notify you if an unknown Bluetooth tracker (including Apple’s AirTags, Tile, Chipolo, and Pebblebee) is moving with you. Google and Apple are working together to develop industry standards for trackers to limit abuse.

The feel-good story of the week is how Twitter user “Bigolas Dickolas, who mostly posts about the anime Trigun, boosted the sales of a science fiction novella (“This is How You Lose the Time War”) originally published in 2019 so it rocketed up the Amazon best seller charts (it reached #3 in overall book sales). This wasn’t influencer marketing. It was a single Tweet saying people should go read this book, with just an image of the cover. There was no link or anything for sale, just people responding how much they love it.

Canada’s draft Online News Act (Bill C-18) has Meta, Google and other platforms concerned. And there’s no denying the hostile intent. Meta notes they were invited to address Canada’s Heritage Committee on the topic “The Response of Companies in the Information Technology Sector to Bill C-18”, but shortly before the hearing it was retitled “Tech Giants’ Current and Ongoing Use of Intimidation and Subversion Tactics to Evade Regulation in Canada and Across the World’”. Google says they are testing “varying amounts of news links” in the search results for a small percentage of Canadian users. The main concern seems to be the requirement to pay news publishers to be able to link to them, with a very broad definition of “news businesses”.

That’s all the updates for this week. Thanks for reading!

Header image background: Graphic created in Canva.