Science & Tech · Cybersecurity · 17 hrs ago
AI could threaten internet encryption before quantum computers, experts warn
Security experts warn AI could weaken some internet encryption before quantum computers become practical.
The concern is that AI may find faster ways to solve hard mathematical problems that protect banking, email and cryptocurrency transactions.
Some encryption relies on calculations that are easy to do in one direction but extremely difficult to reverse.
Researchers are already moving toward post-quantum encryption, designed to resist attacks from quantum computers.
But experts say there is no evidence that AI has broken widely used encryption, and AI could also help improve security.
If a serious weakness is found, changing encryption systems could take years, leaving limited time to respond.
Experts warn that artificial intelligence could threaten some internet encryption before quantum computers become practical.
AI could potentially find faster ways to solve mathematical problems that underpin some encryption, allowing conventional computers to break it.
The warning comes as the security industry is adopting post-quantum cryptography designed to withstand quantum-computer attacks.
There is no evidence that widely used encryption algorithms have already been defeated by AI.
Coinbase cryptography chief Yehuda Lindell said AI could also help create stronger encryption, while warning that the risk remains uncertain.
- Who
- Security experts, including Johns Hopkins cryptographer Matthew Green, warn of a possible risk. Coinbase cryptography chief Yehuda Lindell urges caution about the threat.
- What
- AI could find ways to reduce the calculations needed to break some encryption. No widely used encryption algorithm is known to have been defeated by AI.
- When
- The article was published on November 9, 2026. The warning was reported by The Wall Street Journal on November 7.
- Where
- Not stated
- Why
- Some encryption relies on mathematical problems believed to be too difficult to solve in practice. AI advances could change how difficult those problems are to solve.
AI-risk proponents
AI-risk skeptics
Potential for breakthroughs
AI-risk proponents
Matthew Green says AI progress on difficult mathematical problems raises concern that it could reduce the work needed to break encryption.
AI-risk skeptics
Yehuda Lindell says it is premature to conclude that AI’s mathematical progress will collapse internet security.
Time to respond
AI-risk proponents
Green warns that moving to a new encryption system can take years, leaving too little time after a vulnerability is disclosed.
AI-risk skeptics
Lindell says there is insufficient evidence that AI could defeat even post-quantum cryptography.
Vitalik Buterin called for post-AI cryptography at an Ethereum Foundation annual conference in Mumbai.
The Wall Street Journal reported the warning that AI could threaten encryption before quantum computers become practical.
Hankyung published its article on the potential threat and the differing assessments of experts.
- Publication date
- November 9, 2026
- Risk
- AI could reduce the calculations needed to break some encryption
- Current evidence
- No evidence that widely used encryption algorithms have been defeated by AI
- Post-quantum cryptography
- Designed to withstand attacks by quantum computers
- Conference
- Ethereum Foundation annual conference, Mumbai
1 source
What people are saying
Developers may be told about a vulnerability immediately, but it can take years to switch to a new cryptographic system, leaving too little time to respond.
It is premature to think that AI’s mathematical achievements will collapse internet security.