Google Photos introduces an AI search feature, Ask Photos | TechCrunch
AI

Google Photos introduces an AI search feature, Ask Photos

Comment

Image Credits: TechCrunch

Google Photos is getting an AI infusion with the launch of an experimental feature, Ask Photos, powered by Google’s Gemini AI model. The new feature, which rolls out later this summer, will allow users to search across their Google Photos collection using natural language queries that leverage an AI’s understanding of their photo’s content and other metadata.

While before users could search for specific people, places or things in their photos, thanks to natural language processing, the AI upgrade will make finding the right content more intuitive and less of a manual search process, Google announced Tuesday at its annual Google I/O 2024 developer conference.

For instance, instead of searching for something specific in your photos, such as “Eiffel Tower,” you can now ask the AI to do something much more complex, like find the “best photo from each of the National Parks I visited.” The AI uses a variety of signals to determine what makes the photo the “best” of a given set, including things like lighting, blurriness, lack of background distortion and more. It can then combine that with its understanding of the geolocation of a set of photos or dates to retrieve only those images taken at U.S. National Parks.

Image Credits: Google

This feature builds on the recent launch of Photo Stacks in Google Photos, which groups together near-duplicate photos and uses AI to highlight the best photos in the group. As with Photo Stacks, the aim is to help people find the photos they want as their digital collections grow. More than 6 billion images are uploaded daily to Google Photos, according to Google, to give you an idea of scale.

In addition, the “Ask Photos” feature will allow users to ask questions to get other sorts of helpful answers. Beyond asking for the best photos from a vacation or some other group, users can ask questions that require an almost human-like understanding of what’s in their photos.

For instance, a parent could ask Google Photos what themes they had used for their child’s four last birthday parties, and it could return a simple answer along with photos and videos about the mermaid, princess and unicorn themes that were previously used and when.

Image Credits: Google

This type of query is made possible because Google Photos doesn’t just understand the keywords you’ve entered but also the natural language concepts, like “themed birthday party.” It can also take advantage of the AI’s multimodal abilities to understand if there’s text in a photo that may be relevant to the query.

Another example demoed to the press by CEO Sundar Pichai ahead of today’s Google I/O developer conference showed a user asking the AI to show them their child’s swimming progress. The AI packaged up highlights of photos and videos of the child swimming over time.

Another new feature taps into using search to find answers from text in the photos. That way, you could snap a photo of something you wanted to remember later on — like your license plate or passport number — and then ask the AI to retrieve that information when you needed it.

If the AI ever gets things wrong and you correct it — perhaps flagging a photo that’s not from a birthday party or one you wouldn’t highlight from your vacation — it will remember that response to improve over time. This also means the AI becomes more personalized to you the longer you interact with it.

When you find photos you’re ready to share, the AI can help draft a caption that summarizes the content of the photos. For now, this is a basic summary, which doesn’t offer the option of choosing from different styles, however. (But considering it’s using Gemini under the hood, a smartly written prompt might work to return a certain style if you try it.)

Google says it will have guardrails in place to not respond in certain cases (perhaps no asking the AI for the “best nudes”?). It also didn’t include potentially offensive content when training the model. But the feature is launching as an experiment, so it may need additional controls to be added over time as Google responds to how people put it to use.

The Ask Photos feature will initially be supported in the U.S. in English before rolling out to more markets. It will also only be a text-based feature for now, similar to asking questions of an AI chatbot. Over time, though, it could become integrated more deeply with Gemini running on the device, as on Android.

The company says users’ personal data in Google Photos is not used for ads. Humans also won’t review AI conversations and personal data in Ask Photos, except “in rare cases to address abuse or harm,” Google says. People’s personal data in Google Photos also isn’t used to train any other generative AI product, like Gemini.

We’re launching an AI newsletter! Sign up here to start receiving it in your inboxes on June 5.

Read more about Google I/O 2024 on TechCrunch

More TechCrunch

Struggling EV startup Fisker has laid off hundreds of employees in a bid to stay alive, as it continues to search for funding, a buyout or prepare for bankruptcy. Workers…

Fisker cuts hundreds of workers in bid to keep EV startup alive

Chinese EV manufacturers face a new challenge in their pursuit of U.S. customers: a new House bill that would limit or ban the introduction of their connected vehicles. The bill,…

Chinese EV makers, and their connected vehicles, targeted by new House bill

With the release of iOS 18 later this year, Apple may again borrow ideas third-party apps. This time it’s Arc that could be among those affected.

Is Apple planning to ‘sherlock’ Arc?

TechCrunch Disrupt 2024 will be in San Francisco on October 28–30, and we’re already excited! This is the startup world’s main event, and it’s where you’ll find the knowledge, tools…

Meet Visa, Mercury, Artisan, Golub Capital and more at TC Disrupt 2024

Featured Article

The women in AI making a difference

As a part of a multi-part series, TechCrunch is highlighting women innovators — from academics to policymakers —in the field of AI.

8 hours ago
The women in AI making a difference

Cadillac may seem a bit too traditional to hang its driving cap on EVs. And yet, that hasn’t stopped the GM brand from rolling out — or at least showing…

The Cadillac Optiq EV starts at $54,000 and is designed to hook young hipsters

Ifeel is being offered as part of an employer’s or insurance provider’s healthcare coverage.

Mental health insurance platform ifeel raises a $20 million Series B

Instead of opening the user’s actual browser or a WebView, Custom Tabs let users remain in their app while browsing.

Google Chrome becomes a ‘picture-in-picture’ app

Sanil Chawla remembers the meetings he had with countless artists in college. Those creatives were looking for one thing: sustainable economic infrastructure that could help them scale rather than drown…

Slingshot raises $2.2 million to provide financial services to artists

A startup called Firefly that’s tackling the thorny and growing issue of cloud asset management with an “infrastructure as code” solution has raised $23 million in funding. That comes on…

Firefly forges on after co-founder murdered by Hamas

Mistral, the French AI startup backed by Microsoft and valued at $6 billion, has released its first generative AI model for coding, dubbed Codestral. Like other code-generating models, Codestral is…

Mistral releases Codestral, its first generative AI model for code

Pinterest announced today that it is evolving its Creator Inclusion Fund to now be called the Pinterest Inclusion Fund. Pinterest teamed up with Shopify’s Build Black and Build Native programs…

Pinterest expands its Creator Fund to allow founders

Alex Taub, a longtime founder with multiple exits under his belt, believes it’s time to disrupt the meme industry. “I have this big thesis that meme tech is going to…

This founder says meme tech is the next big thing

Lux, the startup behind popular pro photography app Halide and others, is venturing into video with its latest app launch. On Wednesday, the company announced Kino, a new video capture app…

Kino is a new iPhone app for videographers from the makers of Halide

DevOps startup Harness has shown itself to be an ambitious company, building a broad platform of services while also dabbling in M&A when it made sense to fill in functionality.…

Harness snags Split.io as it goes all in on feature flags and experiments

Microsoft’s Copilot, a generative AI-powered tool that can generate text as well as answer specific questions, is now available as an in-app chatbot on Telegram, the instant messaging app.  Currently…

Microsoft’s Copilot is now on Telegram

HBO’s new documentary, “MoviePass, MovieCrash,” tells a story that many of us know about: how MoviePass, the subscription-based movie ticketing startup, was a catastrophic failure. After a series of mishaps…

MoviePass co-founders speak their truth in HBO’s new documentary 

The watch features a variety of different 3D games, unlocking more play time the more kids move.

Fitbit’s new kid smartwatch is a little Wiimote, a little Tamagotchi

In the video, a crowd is roaring at a packed summer music festival. As a beat starts playing over the speakers, the performer finally walks onstage: It’s the Joker. Clad…

Discord has become an unlikely center for the generative AI boom

After the Wirecard scandal, Germany’s financial regulator BaFin started to look more closely at young fintech startups that wanted to grow at a rapid pace — it’s better to be…

Germany’s financial regulator ends anti-money laundering cap on N26 signups after $10M fine

Among other things, this includes the ability to trace code from source to binary packages across both platforms, single sign-on support and unified project structures.

JFrog and GitHub team up to closely integrate their source code and binary platforms

The company’s public fund disbursement and e-commerce platform makes accepting school tuition and enabling educational enrichment more accessible. 

Tech startup Odyssey goes on journey to help states implement school choice programs

A new startup called Kinnect aims to help people privately save generational memories, traditions, recipes and more. The company’s app, launched this month, lets people create invite-only spaces where they…

Kinnect’s new app aims to help families record and store generational memories

Spotify has hiked its premium subscription in France by an eye-watering €0.13, in response to a new music-streaming tax.

Spotify hikes subscription price in France by 1.2% to match new music-streaming tax

The European Union has taken the wraps off the structure of the new AI Office, the ecosystem-building and oversight body that’s being established under the bloc’s AI Act. The risk-based…

With the EU AI Act incoming this summer, the bloc lays out its plan for AI governance

Solutions by Text, a company that gives people a way to pay their bills and apply for loans via text messaging, has secured $110 million in new growth funding. Edison…

Bootstrapped for over a decade, this Dallas company just secured $110M to help people pay bills by text

Owners of small- and medium-sized businesses check their bank balances daily to make financial decisions. But it’s entrepreneur Yoseph West’s assertion that there’s typically information and functions missing from bank…

Relay raises $32.2 million to help smaller businesses manage their cash flow

When other firms were investing and raising eye-popping sums, Clean Energy Ventures took a different approach. It appears to be paying off.

How Clean Energy Ventures avoided the pandemic bubble and raised a $305M fund

PwC, the management consulting giant, will become OpenAI’s biggest customer to date, covering 100,000 users.

OpenAI signs 100K PwC workers to ChatGPT’s enterprise tier as PwC becomes its first resale partner

Tech enthusiasts and entrepreneurs, the clock is ticking! With just 72 hours remaining until the early-bird ticket deadline for TechCrunch Disrupt 2024, now is the time to secure your spot…

72 hours left of the Disrupt early-bird sale