this post was submitted on 14 Feb 2024
264 points (89.3% liked)
Technology
59587 readers
2472 users here now
This is a most excellent place for technology news and articles.
Our Rules
- Follow the lemmy.world rules.
- Only tech related content.
- Be excellent to each another!
- Mod approved content bots can post up to 10 articles per day.
- Threads asking for personal tech support may be deleted.
- Politics threads may be removed.
- No memes allowed as posts, OK to post as comments.
- Only approved bots from the list below, to ask if your bot can be added please contact us.
- Check for duplicates before posting, duplicates may be removed
Approved Bots
founded 1 year ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
Glad this is being discussed. Having worked adjacent to the authentication market, I have mixed feelings about it, though.
There are a few problems with passkeys, but the biggest one is that no matter what, you will always need a fallback. Yes, Apple promises a cloud redundancy so you can still log in even if you lose every device.
But that's just Apple's ecosystem. Which, for what its worth, is still evolving. So the passkey itself is phishing-resistant, but humans still aren't. Fallbacks are always the weakest link, and the first target for bad actors. Email, or sometimes phone and SMS, are especially vulnerable.
Passkeys in their current iteration are "better" than passwords only in that they offload the fallback security to your email provider. Meanwhile, SIM swapping is relatively ready easy for a determined social engineer, and mobile carriers have minimal safeguards against it.
Usability? Great, better than knowledge-only authentication. Security? Not actually that much better as long as a parallel password, email, or SMS can be used as a recovery or fallback mechanism.
I'm not saying passkeys are bad, but I'm tired of the marketing overstating the security of the thing. Yes, it's much more user-friendly. No one can remember reasonably complex passwords for all 100 of their online accounts. But selling this to the average consumer as a dramatic security upgrade, especially when so many still run passwords in parallel or fall back to exploitable channels, is deceptive at best.
So you really need two independent devices with their own passkeys to back each other up.
Not sure exactly what you're getting at, but any authentication model must be designed with the assumption that a user can lose all their devices, passkeys included. That's where fallbacks come into play. Even with Apple's system, you can recover your keychain through iCloud Keychain escrow, which (according to their help page) uses SMS:
While SIM swaps aren't super common, they're not the most difficult attack. Passkeys are strong against direct attacks, for sure. But if I can reset your account using a text message sent to a device I control, is it really that much more secure?
So if you lose access to all of your devices, you're completely locked out of everything until you're able to get a new working phone activated on a trusted phone number? The trade-off of inconvenience for security here just doesn't seem worth it to me.
Depends on the provider in question. While Apple does allow SMS recovery, they also let you designate a trusted contact who can let you in as an alternative. This is obviously more convenient (if you have a friend or family member who can be available when you need them), but the situation with SMS vulnerabilities is still my main gripe.
Is it possible to use some kind of fingerprinting to identify people? It works for marketers, could that idea be used for security?
I am a total noob who is interested, if I come across as uninformed it is because I am.
Totally! Browser and device fingerprinting are commonly used as first-line defenses against ATOs (account takeovers). There are other kinds of fingerprinting, like those that can learn about your installed hardware and drivers. Really, I'm learning about more fingerprinting methods all the time. That said, decisions are usually made based on several different information sources. These include variables like:
There's even some buzz around "behavioral biometrics" to identify individuals by how they type, but this is still not the sole method of identification. It's mainly about flagging bots who don't type like humans. However, learning how an individual types can help you determine if a subsequent visitor is the actual account owner or a bad actor.
In my experience, fingerprinting and adjacent identity proofs are rarely used in isolation. They're often employed for step-up authentication. That means if something doesn't match up, you get hit with a 2FA/MFA prompt.
Step-up can be pretty complex if you want it to be, though, with tons of cogs and gears in the background making real-time adjustments. Like you might not even realize you've been restricted during a session when you log in to your bank account, but once you try to make a transfer, you'll get an MFA prompt. That's the UX people in action, trying to minimize friction while maintaining security.