Sun 28th Jun 2026
Quantum Key Distribution - part 4
Service: Patents
Sectors: AI and data science
In part four of our Quantum Key Distribution series, Aunt Carol enters the picture! Rachel Barrett explains about her entangled photons, and the fine print.
Previously, we caught Steve by noticing he couldn’t touch photons without leaving smudges. Today, we catch him a different way: by checking whether the universe is still doing that spooky correlation thing it promised.
BB84 is the ‘prepare and measure’ flavour of QKD: you prepare photons in particular states and Bertha measures them. But there’s another option, if you fancy bringing a third party into your life, and who doesn’t?
Enter Aunt Carol.
Aunt Carol’s entangled delivery service
Instead of you sending photons to Bertha, Aunt Carol sends entangled pairs of photons: one photon from each pair to you, and the other to Bertha.
These aren’t just any photons. They are entangled in such a way that, when you and Bertha measure them in the same basis, your outcomes are strongly correlated (often anti-correlated, depending on the exact entangled state). This means you can generate shared bits without trusting that any single photon ‘carried’ a pre-chosen value from you to her. The shared randomness emerges from measurement outcomes that only line up in the right way if the entanglement is intact.
Aunt Carol, for her part, is delighted. She feels useful. She has a purpose. She also insists on bringing snacks, which is, frankly, the best part.
The measurement game (now with extra bases)
You and Bertha independently choose measurement bases at random for each photon you receive. In many entanglement-based schemes, you pick from more than two measurement settings. This sounds like extra effort, but it’s the good kind - the kind that makes Steve’s life worse.
You each measure your photons and record outcomes as bits.
Then, as before, you talk over a classical channel and announce which bases you used (not the outcomes). Whenever you happened to choose compatible bases, you keep those results as candidate key bits.
If the state you’re using gives anti-correlated results, that’s fine: one of you flips your bits, and then you match. You do it. You’re more reliable.
Catching Steve: correlations as a tripwire
Now for the good part.
If the photons are genuinely entangled and nobody has meddled, the measurement outcomes exhibit correlations that are stronger than anything Steve can reproduce by intercepting, measuring, and re-sending classical lookalikes.
So you use some of the data you didn’t keep for the key - the ‘discarded’ bits where you and Bertha measured in different bases - to compute a correlation statistic. If the correlations are sufficiently strong, it’s consistent with entanglement being present. If they’re not, something has disturbed the system.
And who, in this story, loves disturbing systems?
Steve.
His interference tends to break entanglement (or, more precisely, degrade the non-classical correlations you’re looking for). That means even if he’s careful, he can’t quietly keep the same ‘spooky’ pattern intact while also learning your bits. The universe is, once again, being uncooperative towards his hobbies.
BB84 vs entanglement-based QKD (a quick compare)
BB84: you prepare states; Bertha measures.
Security comes from measurement disturbance + basis mismatch causing detectable errors.
Entanglement-based: a source distributes pairs; both of you measure.
Security is phrased in terms of correlations: meddling degrades them.
Both ultimately rely on the same idea: Steve can’t learn the key without leaving evidence.
The fine print (kept mercifully short)
A few things that are true, even if they are less fun than cake:
1. You still need authentication.
QKD detects eavesdropping, but it doesn’t automatically prevent Steve from impersonating you to Bertha and Bertha to you. That’s why an authenticated classical channel (or bootstrapped trust) is part of any real deployment.
2. Noise is real.
Perfect correlations are a fantasy. QKD is built to tolerate some error and still squeeze out a secure key via privacy amplification.
3. QKD gives you keys, not magically secure everything.
You still use conventional symmetric crypto to encrypt your messages. QKD is the key-establishment layer — a very fancy one — not the whole communications stack.
And that’s it.
Steve can be hardy. Steve can be patient. Steve can even buy a quantum computer. He just can’t be quiet.
You’re now free to share your most heinous secrets with Bertha without risking Steve discovering the terrible truth about you.
You’re welcome.
For more information, please contact Tom Mahon
This briefing is for general information purposes only and should not be used as a substitute for legal advice relating to your particular circumstances. We can discuss specific issues and facts on an individual basis. Please note that the law may have changed since the day this was first published in June 2026.


