Most consumer digital service providers assume that if someone interacts with them using a particular digital key, they must be who they say they are. In the purist view of crypto, this artifice has been dropped, and interactions occur between keys – regardless of who or what is holding the key. Neither approach affords a natural bridge for people living at the intersection of the digital and physical worlds.
This week I signed up for a new digital service and at the end of the registration, the provider sent an SMS with a code to the phone number I’d provided. I entered the code on the provider’s website and got the message – “that’s great, now we know it’s you, we can get started”. They “know it’s me” because the inputs and outputs on the other side of their digital interface satisfied their software rules. All they know is that there’s another piece of software that has responded to their interface – and this sets the scene for the way they think of their customers.
Physical key technology has served us pretty well for a few thousand years but does have some well-known drawbacks – they are easily copied or stolen. You might not know which of your friends (or friends friend’s) have been lent keys. You have to change the locks and keys when a lock is compromised. Over time you end up with quite a collection of keys on your key ring – sometimes forgetting what they were for etc.
In most situations, we don’t think of a physical key as a way to authenticate who we are. Just because someone opens my front door with a key doesn’t mean that I trust them without question to have open access to the house. Imagine being blindfolded with earplugs in and assuming that anyone with your front door key must be trustworthy. Yet that is precisely what many digital service providers are doing – not because they think you’ll ransack their website – but because they don’t put a value on the real human being that holds the key.
We have willingly entered into a dysfunctional arrangement whereby we do all the work to manage an ever-increasing number of unique digital keys and door addresses in order to access services where the provider doesn’t value our individuality.
Digital keys come in many different forms; passwords, SMS message codes, codes generated by apps authenticated by other keys and more recently cryptographically verifiable codes.
We all knew that the password key approach became a mistake as soon as we needed to open more than seven plus or minus three digital doors. For many of us, our digital “key ring” has hundreds of keys, most of which look the same and we spend a lot of time fumbling in the dark, trying to find the key that fits the lock to access a digital service. If you try the wrong key too many times, you have to change your own lock and cut yourself a new key – remembering to remove the old key from the key ring.
Authentication that uses a phone number like SMS codes assumes that it must be you if a digital response is received from a digital request. This is just lazy. The much-vaunted ‘two factor authentication’ approach is only useful if the two factors use different contexts, including, ideally, something that is directly connected to a human experience.
Biometric device-based digital key mechanisms offer some improvement by assuming that if you can access your device, then the device maker can be trusted to know it must be you – or at least someone who has your finger or face. None the less, until we get the implants, it’s impractical to require you to have your phone or watch at all times to access a digital service.
Cryptographic keys cut the human out entirely and simply trust the computation of the input and output. Being “digitally native”, they also have the additional functionality of being a piece of software that can be used in association with other algorithm software to encode and decode digital content. Kind of like a physical key that changes a door into a window when turned in a lock.
The “beauty” of cryptographic keys and associated safes is that, for all practical purposes, they resist all currently known safe cracking techniques and don’t require any other human intervention – the pieces of software are either compatible or they aren’t. If they are, then the key opens the lock. This is great for all of you algos out there – not so compatible with a society predicated on exchanges between humans.
So how do you keep your digital keys safe? You put them in another software or even a physical box with a different key. You might make this box hard to find or access. Maybe put it in a bank vault where you will need many more forms of identity than just holding a key.
Like a physical key, anyone with a copy of a digital key can open the lock. Unlike a physical key, if you lose a digital key (and all copies) the lock will never open again.
Does this really sound like a huge leap forward for civilisation?
Like the other themes in these posts, digital authentication approaches are failing because they expect people to operate interfaces designed for digital things. And like other themes, new digital approaches tend to carry forward design concepts from the physical world without appreciating the unique aspects of a blended digital and physical environment.
To live happily and securely in a blended physical and digital world, we must adopt new attitudes and blended approaches for authentication and trust. These new approaches won’t succeed by requiring service consumers to become cryptographic locksmiths. The only choice is for us to foster a deeper and more secure relationship with a trusted digital gatekeeper designed to serve human consumers and participate in digital-to-digital interactions on our behalf.