Microsoft is moving antivirus programs from running at kernel level

Antivirus software will soon be moved out of the kernel mode in Windows. This change is part of Microsoft's Windows Resiliency Initiative (WRI).
Last year, millions of Windows PCs crashed with a blue screen due to a faulty update for Crowdstrike. In the aftermath of the incident, Microsoft held a security summit with the intention to prevent such issues in the future. Several security vendors, including Bitdefender, CrowdStrike, ESET, SentinelOne, Trellix, Trend Micro, and WithSecure, joined the Microsoft Virus Initiative (MVI) 3.0 program to collaborate with Microsoft and improve the security and reliability of Windows.
Microsoft says that it will release a private preview of the Windows endpoint security platform to its MVI partners. The changes will require antivirus software, and endpoint detection and response (EDR) apps, to run in user mode like most apps do. Microsoft highlights that running apps with administrator permissions opens the door to malware, which could infect a user's computer, and wreak havoc on critical system resources, causing disruptions, data loss, etc. This was what had caused the Crowdstrike BSODs last year.
Security vendors will be able to test their software, and request changes if required, to ensure that their antivirus products run fine in user mode. The Verge quotes David Weston, vice president of enterprise and OS security at Microsoft, who said that "We’re not here to tell them how the API should work, we’re here to listen and provide the security and reliability".
Instead of laying down the rules, Microsoft is collaborating with antivirus vendors to share feedback and co-engineer the system. It could take a while to get things sorted with the previews, but this is a nice approach.
Weston also hinted that these changes could also affect kernel-level anti-cheat mechanisms used in games, Easy Anti-Cheat for example. It could be a thing of the past, which is welcome news.
“Microsoft highlights that running apps with administrator permissions opens the door to malware, which could infect a user’s computer, and wreak havoc on critical system resources, causing disruptions, data loss, etc. This was what had caused the Crowdstrike BSODs last year.”
This is not what caused the Crowdstrike BSODs. This was caused by software running as a Windows driver in ring 0, with full access to the Windows kernel.
Administrator privileges is a separate user-mode concept and refers to standard applications that run under the Administrators group or otherwise elevated.
There is only one reason why M$ is willing to play well with others in the industry, To suit their nefarious goals.
M$ played well with Linux, now they are all over the Linux kernel…including the network, there is no getting away from them.
Again, the user ends up being the Loser. **** MS
Any time I want to run a scan on a Windows box (kind of a joke anyway since Windows is a rootkit with rootkits) I boot into a Linux Live environment and scan away, without Windows being directly loaded.
That being said, I wouldn’t trust Windows on ANY hardware. It’s proprietary, the company is a convicted monopoly which is above the law, hasn’t been broken up as they should’ve (yet Google is supposed to be broken up?) been and is outright criminal!
“We’re not here to tell them how the API should work, we’re here to listen and provide the security and reliability”. From a company that cannot get M$ 365 or Outlook right.
ESET will get it right! I have more confidence in them than M$.
Compromise free for 25 years, I have.
This basically sounds like a terrible idea. Once the process does no longer run in ring 0, but in ring 3 .,. there is no protection of any kind from malicious software that managed to elevate to admin privileges anymore. Running the processes as system worked pretty well (which is why many cybercriminals switched to ransomware in the first place, so they wouldn’t have to circumvent this in the first place)
Yes, this change WILL increase system stability, no doubt, and crashes like the Crowdstrike incident won’t happen again, but the AV process will no longer be properly protected from other processes and as a result be more or less useless. While the usefulness of signature-based AV shouldn’t be overemphasized, no AV (which is identical to an AV running in user mode) is a terrible idea, as everybody knows who has ever manager a larger organization with a few thousand or more nodes.
So this is absolutely terrible news. I’m not quite sure what’s going on I’m MS mind, if they think, that throwing security in favor of a perceived stability increase out of the window is going to make their system more useful for professional use.
The Crowdstrike problem was, for once, not a Windows issue, but a problem of a) a software vendor not having properly tested its patches (which is something that was bound to happen, so no big surprise here. That was an event, which was to be expected. Risk management anyone ? ) AND especially b) a vast number of companies allowing to run untested code into their productive networks. I mean, that’s what you have development and quality assurance networks for, right ? You only let changes enter the production environment, once it has been proven in the QA network to be benign. YOU DO NOT roll out untested changes into the production network, because that may endanger it’s stability and reliability.
THAT was the real issue. If even many Fortune 500 companies do no longer follow trusted, time proven and working security rules why is everyone surprise if bad things happen ?
Unfortunately, there is no working patch for human stupidity. This means MS can change the architecture as much as they want, if people don’t understand or don’t care how to properly run an IT architecture in a professional way, even if that’s exactly their job and what they are paid for, then it won’t get better, just worse.
Reading things like this actually makes Idiocracy sound like a predictive documentary, and not an unfunny movie.
Just out of curiosity, has anyone ever had a 3rd-party antivirus app identify and stop dangerous malware that plain old boring MS Defender wouldn’t have?
Why do these apps even exist in 2025? This functionality has been baked-in to the OS for years now. The fact that a few 3rd-party apps supposedly test 3% better on some academic AV-Comparitives test or whatever is close to meaningless in the real world. Any half-assed malware threat is run though VirusTotal to make sure it doesn’t raise any flags before it gets released into the wild.
Kaspersky did stop the “WannaCry” ransomware in 2015, while other antiviruses, including MS Defender, didn’t. It’s the only third party app worth installing (only paid version, the free one sucks), if you don’t mind being spied by Russians ofc.
While Defender is nice and has the additional advantage of not tearing any new holes into the system’s attack surface, it has a massive disadvantage: it doesn’t have a proper management environment for centrally managing and integrating thousands of AV clients, meaning: you can’t use it in a professional environment. So for personal use, it’s fine for professional use: it ain’t.