Going Dark, and the era of law enforcement hacking
Going Dark, and the era of law enforcement hacking
Author Profile: Matthew Green is a professor at Johns Hopkins University and a specialist in cryptography. His career involves the design and scrutiny of cryptographic frameworks used in digital content protection, payment systems, and wireless networks. His primary research focus is the intersection of cryptography and user privacy.

Reflections from Baltimore
I have recently returned from attending Usenix Security, which took place in my home city of Baltimore. My time there was largely spent in two types of discussions:
- Technical exchanges with peers.
- Explaining to my colleagues that Baltimore is
exactly like the show The Wirenot actually like The Wire.
While I harbor various anxieties regarding the impact of Artificial Intelligence on our professional landscape (depending on how you define "the field"), there is one specific, somewhat paradoxical concern I want to highlight: I fear that AI will make software excessively secure.
The Paradox of "Too Much" Security
At first glance, "too secure" sounds like a victory. However, there is a significant downside. Specifically, I am worried that U.S. law enforcement and intelligence agencies are on the verge of "going dark."
Definition: "Going Dark" The phenomenon where government agencies suddenly lose a massive portion of their surveillance and investigative capabilities due to technological barriers.
This isn't just a crisis for the agencies; it is a problem for everyone who advocates for general computer privacy and security.
A History of Surveillance Evolution
To understand the current predicament, we have to look at the trajectory of electronic surveillance. The Wire (specifically the first season) serves as a perfect time capsule for the state of surveillance circa 2002.
The Old Guard (Early 2000s)
In that era, police monitored drug traffickers who relied on:
PayphonesBurner phones
From a technical standpoint, a detective from 1989 would have felt right at home in 2002. However, within a decade, these methods became obsolete.
The Shift to Data (Late 2000s - 2016)
The landscape shifted with the proliferation of smartphones and SMS. Because these devices store data rather than just transmitting it, they became goldmines for law enforcement. This "golden age" of access was short-lived.
Timeline of Encryption Adoption
| Year | Event | Impact |
|---|---|---|
| 2010 | Apple encrypts iPhone storage | Keys are derived from user passcodes: |
| 2011 | iMessage E2EE | End-to-end encryption (E2EE) for texts |
| 2014 | WhatsApp Growth | Reaches 600 million users |
| 2016 | WhatsApp E2EE | Nearly 1 billion users now using default E2EE |
The Surveillance Transition Flow
The Institutional Response
Law enforcement agencies did not watch this transition happen in silence.
- 2014: FBI Director James Comey launched the "Going Dark" initiative.
- Goal: Start a "national conversation" to pressure providers into making communications legible to counterintelligence and police.
- Outcome: By 2016, the agency transitioned from dialogue to a more aggressive strategy.
Technical Implementation of Privacy
The shift can be summarized by the move from simple transmission to complex cryptographic engineering. For example, the transition from a cleartext message to an encrypted one looks like this in concept:
# Conceptual shift in messaging
def send_message(text, recipient):
# OLD WAY:
# transmit(text)
# NEW WAY:
encrypted_text = encrypt(text, recipient.public_key)
transmit(encrypted_text)
The result is a world where the "wire" is no longer something that can be simply tapped.

Further Reading & Resources:
- Cryptopals challenges
- Applied Cryptography
- Journal of Cryptographic Engineering
