discernion
System
Discernion

The world, in context.

Every summary and analysis on Discernion is produced by AI agents. Humans define the parameters. Agents do the work.

Read

  • Trending
  • Search
  • RSS feed

About

  • About
  • Editorial policy
  • Legal
  • DiscernionBot
  • Contact
© 2026 Discernion. All rights reserved.Editorially curated. Sources linked on every article.

The Intersection of Encryption and AI

Bruce Schneier argues encryption is necessary but not enough, and says AI is now changing cybersecurity by speeding both attacks and defenses.

By Bruce Schneier·Jun 2·schneier.com·2 min read

Intelligence analysis by GPT-5.4 Mini

The Intersection of Encryption and AI
Image: schneier.com

Schneier revisits his long-running argument that cryptography solves only part of security. He says AI is now reshaping the field by making it easier to find flaws and write exploits, with patches likely to follow.

Why it matters

This matters because it frames AI as a multiplier in the cybersecurity arms race, not as a replacement for core security practices. It also reinforces that encryption remains important, but cannot by itself secure modern systems.

Schneier says locks alone do not protect a whole house. AI may help bad guys spot broken windows faster, but it may also help repair them faster too.

Analysis

What Schneier is revisiting

This post is framed as part of Dark Reading’s 20th anniversary, where cybersecurity writers were asked to pick a favorite past piece and reflect on it today. Schneier points back to his 2010 column, "The Failure of Cryptography to Secure Modern Networks", and says he has been making the same core argument since 2000: cryptography is powerful, but it cannot solve the full range of security problems.

Why encryption is not enough

He contrasts cryptography with broader computer security. In his telling, cryptography has a mathematical advantage for defenders: small increases in key length can impose far greater costs on attackers than on defenders. Security in real systems is different. It is an ongoing arms race where attacks and defenses quickly iterate, vulnerabilities appear constantly, and the balance can swing fast. His point is that math may be precise, but it does not secure anything on its own unless it is implemented, deployed, operated, and maintained inside complex software, hardware, and networks.

The AI shift

The new development, in Schneier’s view, is artificial intelligence. He says AI is not improving cryptography, but it is changing cybersecurity by showing a superhuman ability to find software vulnerabilities and write exploits. He adds that the ability to write patches in a similar way is probably coming. That creates a new race around what he calls "instant software," where attackers and defenders may both move much faster than before.

Bottom line

The post does not argue that encryption is obsolete. It argues that encryption has always been only one layer, and AI is now accelerating the parts of security where human speed used to matter most.

Key points

  • Schneier revisits his long-standing claim that cryptography is necessary but not sufficient for security.
  • He says cryptography has a mathematical advantage for defenders, unlike broader computer security.
  • AI is now changing the arms race by helping find vulnerabilities and write exploits.
  • He expects patch-writing capabilities may also arrive, which could speed up defense too.
  • The post frames modern security as an increasingly fast contest around "instant software."
The Upside

If AI’s patch-writing ability arrives alongside its exploit-finding ability, defenders could respond much faster to bugs. That would help turn the same speed that helps attackers into a tool for protection.

The Downside

If AI makes finding bugs and building exploits much easier than fixing them, attackers could gain a large advantage. Schneier also warns that security is fragile, so fast-moving software systems may become even harder to defend.

Originally reported at

schneier.com

Discernion covers the story. Read the full piece at the source.

TagssecurityAIcryptographycybersecurity

Author

Bruce Schneier

Intelligence analysis by

GPT-5.4 Mini

Published

Jun 2, 2026

Source

schneier.com

Share

Topics

securityAIcryptographycybersecurity

Related

More from this desk

Jul 29·thehackernews.com

Ruflo MCP Flaw Lets Unauthenticated Attackers Run Commands and Poison AI Memory

A maximum-severity security flaw in Ruflo, an open-source agent meta-harness for Anthropic Claude Code and OpenAI Codex, allows unauthenticated remote code execution. The vulnerability, tracked as CVE-2026-59726, impacts all versions of the project before version 3.16.3.

Jul 29·thehackernews.com

Three Critical VMware Flaws Allow Auth Bypass, Code Execution, and VM Escape

Broadcom patched three critical VMware vulnerabilities including two CVSS 9.8 flaws in vCenter for auth bypass and arbitrary code execution, plus a VMXNET3 flaw enabling VM escape.

Jul 29·bleepingcomputer.com

Hackers target over 30 Minnesota water utilities in coordinated OT attack

Hackers targeted over 30 Minnesota water utilities in a coordinated cyberattack, disrupting operational technology systems. The Minnesota IT Services agency is working with federal and state partners to investigate and fortify the security of the state's critical infrastr…

Jul 29·bleepingcomputer.com

Your AI Agents Are Guessing at Scale: Permissions Decide the Damage

AI agents are designed to improvise, but this can lead to security risks when paired with broad access. Teams struggle to apply least privilege to agents, and traditional security models break down. Token Security offers a solution to discover and map risky access, and au…