Microsoft Modifies Recall AI Feature Amid Privacy, Security Failings

  /     /     /  
Publicated : 23/11/2024   Category : security


Microsoft Modifies Recall AI Feature Amid Privacy, Security Failings


In response to recent public outcry, Recall is getting new security accouterments. Will that be enough to quell concerns?



Microsoft is adding new security measures to assuage widely publicized concerns over its new Recall AI feature. Some, though, still arent convinced the company went far enough.
Its now just eight days until Microsoft releases Recall, a new artificial intelligence (AI)-driven program that will periodically take, store, and analyze screenshots of Copilot+ PCs as theyre being used day-to-day. Recall is supposed to act like a kind of memory bank, allowing users to instantly find and reference things theyve come across recently: apps, websites, images, and documents.
From the outset, Recall has been criticized as a potential
goldmine for personal data theft
. The noise got loud enough that, on Friday, Microsoft announced
three new security-oriented updates
for it:
In a reversal of its initial stance, Microsoft will now ship Recall turned off by default.
Users will need to enroll in
Windows Hello
in order to enable it, and so-called proof of presence will be required to use its primary features.
Recall data will be encrypted, and only decrypted and accessible once a user authenticates via Windows Hello.
Though they may represent a step in the right direction, experts remain skeptical that these changes will be enough to protect users most sensitive passwords, photos, personally identifying information (PII), and financial information from hackers.
Many security experts cringed when Recall was announced, but few more than Marc-André Moreau, CTO of Devolutions. He worried that Windows newest toy would inevitably capture and store visible passwords from his companys software for managing remote connections. With such passwords in hand, hackers would be able to easily connect to and manipulate any victim PC.
Looking at documentation for how Recall works, he recalls, it literally said that it wouldnt make an effort of removing sensitive information, credentials, or PII — anything which you would want scraped out, it would just keep in local files.
Microsofts logic, it seemed, was that because Recall screenshots were stored only on the users machine, they would remain safe from remote access. Microsoft has this new chip which makes it possible to do the processing locally, and they thought that everybody would be fine since the data isnt uploaded to the cloud, Moreau explains. But you wouldnt install a keylogger on your machine just because the files are stored locally. Files can be grabbed by malware. So why would you enable Recall?
To demonstrate the point, he performed a simple red team exercise. In his telling, I didnt have to do much. I just set up an environment, used some tool that somebody made online to force-install it, and then I installed [Devolutions] Remote Desktop Manager. I clicked view password, then record, and then I found the database. I opened it, and I could see the extracted password alongside the screenshot that includes the password.
Other researchers have also found simple ways of
accessing sensitive data in Recall screenshots
. One has already developed and released an
open source tool
to help speed up the job.
To try to protect his customers, Moreau next looked for a way to exclude his companys software from Recall by default. He came up short.
Users will have more control over their data privacy now, thanks to Microsofts turning off Recall by default.
Moreau is skeptical, though, that Windows Hello can be fully and properly integrated into Recall without delaying its preview release, which is mere days away. Im in software, things dont happen that fast, he says.
Dark Reading reached out to Microsoft for comment on how it will be able to marry Windows Hello and Recall in time for June 18. In response, Microsoft said in a statement: As we shared in our 
May 3 blog
, security is our top priority at Microsoft, in line with our ⁠
Secure Future Initiative
 (SFI), and we are evaluating Recall through that lens. As we implement SFI across Microsoft, we may shift some feature release dates and will update our public roadmaps as this happens.
In the barely a month since that blog post, and Satya Nadellas letter
prioritizing security above all else
, for some critics, Recall recalls other AI products that are getting rushed to market.
Ironically, AI could well solve these programs most pressing security flaws. I could upload a Recall screenshot to ChatGPT today and tell it to identify the data which looks sensitive, and it will be able to, Moreau notes. They could have used their AI chip to help solve this [data leakage] but they didnt even try. They were too eager to ship.

Last News

▸ Hack Your Hotel Room ◂
Discovered: 23/12/2024
Category: security

▸ Website hacks happened during World Cup final. ◂
Discovered: 23/12/2024
Category: security

▸ Criminal Possession of Government-Grade Stealth Malware ◂
Discovered: 23/12/2024
Category: security


Cyber Security Categories
Google Dorks Database
Exploits Vulnerability
Exploit Shellcodes

CVE List
Tools/Apps
News/Aarticles

Phishing Database
Deepfake Detection
Trends/Statistics & Live Infos



Tags:
Microsoft Modifies Recall AI Feature Amid Privacy, Security Failings