All About Cookies is an independent, advertising-supported website. Some of the offers that appear on this site are from third-party advertisers from which All About Cookies receives compensation. This compensation may impact how and where products appear on this site (including, for example, the order in which they appear).
All About Cookies does not include all financial or credit offers that might be available to consumers nor do we include all companies or all available products. Information is accurate as of the publishing date and has not been provided or endorsed by the advertiser.
The All About Cookies editorial team strives to provide accurate, in-depth information and reviews to help you, our reader, make online privacy decisions with confidence. Here's what you can expect from us:
When was the last time you scrolled all the way back through your camera roll? Somewhere in your extensive gallery are your vacation photos, your kids’ birthday parties, a screenshot of last month’s bank statement, a photo of your prescription, and maybe even a picture of your passport you took “just in case.”
Now imagine handing all of this to an AI model. That’s exactly what Dazzle, a new AI assistant, is asking you to do. In return, it promises to know you better than you know yourself. It can shop for you, plan your vacation, and fix things even before you ask.[1]
It might look like a tempting offer, akin to having a Tony Stark-like Jarvis for yourself, but before you tap “Allow Full Access,” it’s worth understanding just how privacy-invasive and potentially dangerous this can be.
Your camera roll can reveal where you go, who you know, what you like, and even sensitive documents, while a breach could put that information in the hands of cybercriminals for targeted advertising, phishing, or identity theft.
Here’s everything you should know about this new AI tool and the steps you can take to protect your privacy.
Dazzle, built by ex-Google executive and former Yahoo CEO Marissa Mayer, is a new personal AI assistant built around your camera roll. Dazzle AI raised $8 million in seed funding in December 2025.
So far, AI chatbots have required prompts to get any work done. AI has little context about you to begin with, which can make it harder to understand what you need. Dazzle instead reads your photos, which can reveal a lot of personal details about you, including your tastes, hobbies, travel details, people close to you or your acquaintances, and even what you own.
Dazzle treats every photo as a possible request: Why did you take the photo, and does it mean something needs to be bought, fixed, or scheduled? Dazzle comes with various features that make this possible:
As “dazzling” and futuristic as Dazzle might sound, it comes with a lot of privacy questions.
For starters, a camera roll isn’t just pictures; it can be a detailed record of your life, including your home and work locations, routines, relationships, hobbies, purchases, and the people close to you.
Giving an AI tool full access to your photo gallery means giving it access to a huge amount of personal context about your life and the people you share it with. Dazzle itself says its photo intelligence can understand your style, interests, hobbies, travel, the people who matter to you, and what matters to them.
Photos can also carry metadata, including GPS coordinates and timestamps, which can reveal where you live, work, or travel. Screenshots are an even bigger concern. Your camera roll can contain screenshots of bank balances, OTPs, ID documents, medical reports, passwords, and private chats.
Dazzle specifically treats photos and screenshots as potential prompts, so giving it full access could mean letting it analyze information you may never have intended to share with an AI assistant.
Such a rich collection of personal data would also make an AI company a lucrative target for hackers. We’ve already seen how damaging a breach can be when a service stores large amounts of image-related data.
Recently, Gyazo suffered a data breach in which attackers accessed 23.62 million user records and around 490 million image metadata records, including EXIF location data, OCR text, and hashed passphrases for private images.
If an AI assistant holding a cache of personal information were breached, criminals could use the stolen data to launch highly targeted phishing scams, commit synthetic identity theft, or attempt financial fraud.
Dazzle says it automatically excludes photos that it suspects are sensitive. However, it doesn’t clearly define what it considers "sensitive," and what happens if the system fails to recognize highly private images as sensitive.
On top of all that, AI agents don’t just observe; they can act. That moves the question from “What does it know?” to “What can it do?” AI agents have already demonstrated that they can sometimes act beyond what a user intended.
In one recent case, a user asked an AI agent to book a gym class, but the agent exploited an authorization flaw in the gym’s booking system to access unavailable classes and even bump the user up the waitlist by removing another person’s reservation.
A tool like Dazzle also puts other people at risk, since your photos often contain friends, family, co-workers, or strangers who didn’t consent to being analyzed by an AI tool. It’s almost exactly the kind of privacy concern that has fueled backlash against Meta’s smart glasses, including bans in some public places.
If you’re curious to try Dazzle or a similar AI assistant, you don’t have to hand over your entire life history. Here are a few simple steps that can significantly reduce the risk.
Dazzle offers a glimpse of where personal AI could be heading, with assistants built to understand your life and serve as an intuitive helper from the photos you take rather than waiting for you to type out a prompt.
But that convenience comes with a trade-off. The more an AI knows about you, the more sensitive information you’re trusting it to handle. A data breach at the AI company or an AI agent going rogue could have drastic consequences for your privacy, including letting bad actors profile you and target you with cyberattacks.
If you’re trying these AI tools, only share the photos you’re comfortable handing over, and avoid giving them unrestricted access to your entire camera roll. Also, make it a habit to regularly clean up sensitive screenshots.
Credit monitoring service with 3-bureau scores updated monthly and a breakdown of what's affecting your score
Data removal covers 425+ broker sites, plus a digital account cleanup that flags risky accounts in your inbox
Identity theft insurance up to $1M for individuals, with 24/7 phone and email support
/images/2026/07/20/aura_data_removal_review.jpg)
/images/2023/10/27/is-chatgpt-safe.jpeg)
/images/2026/10/01/happy_group_and_selfie_with_friends_collage_and_picture_with_personal__bWoeSMs.jpg)
/images/2026/09/29/copilot-photos-reviewed-by-strangers.jpeg)
/images/2026/09/28/computer_screen_with_chatgpt_website_and_person_using_mobile_phone_to_C3CsP2k.jpeg)
/images/2026/09/21/chatgpts-hidden-cookie.jpeg)
/images/2026/09/18/the_hidden_risks_in_anthropics_new_claude_money_feature.png)
/images/2026/09/17/assorted_ai_apps_are_seen_on_an_iphone_including_chatgpt_claude_perple_eqiNhV2.jpg)
/authors/krishi_chowdhary_headshot.jpg){kind=small}
/authors/kate-quinlan-new.jpg){kind=small}
/images/2024/05/17/7249df62-b35e-4ac5-bfc4-0e28a6a30b9e.png)
/authors/krishi_chowdhary_headshot.jpg)