How AI could make it harder for governments to use hacking tools
Cryptography professor Matthew Green has sparked debate in the cybersecurity community by suggesting that AI could make software so secure that governments lose their ability to lawfully hack suspects. His argument is that as AI-powered tools become better at finding and patching software vulnerabilities, the "uneasy truce" under which law enforcement uses hacking tools and spyware instead of demanding encryption backdoors could break down, potentially prompting governments to push for backdoors once again.
Green's warning follows years of "going dark" debates, dating back to 2014 when then-FBI director James Comey said encryption in apps like Signal, WhatsApp and iMessage was hampering investigations. Since then, agencies have relied on purchased exploits and spyware rather than backdoors to access devices. TechCrunch canvassed several security professionals, including researcher Luna Tong and Crowdfense CTO Paolo Stagno, finding mixed views: some agree AI-driven bug-hunting is causing a temporary "gold rush" that will make flaws scarcer, while others doubt states would ever abandon surveillance capabilities regardless of how AI reshapes the vulnerability landscape.
- Cryptographer Matthew Green warns AI could make software too secure to hack
- Governments have relied on exploits and spyware instead of backdoors since 2014
- Security experts are divided on whether AI will truly eliminate exploitable bugs
New here? Start with this
End-to-end encryption protects messages so that only the people in a conversation can read them; the service provider cannot routinely hand over their contents. Governments have argued that this can make it harder to investigate serious crime, a debate often called “going dark”.
Rather than requiring companies to build a way around encryption, investigators have also used hacking tools to exploit flaws in phones and other software. These tools can give access to a suspect’s device, but they depend on weaknesses that developers may later fix. The key players include governments seeking evidence, technology companies securing their products and researchers who find software flaws.
Artificial intelligence may help find and repair those flaws more quickly, though how much it will change access for investigators is disputed. The issue matters because it could affect the balance between secure digital services, criminal investigations and government surveillance powers.
Both sides, in good faith
The strongest fair case each way — we don't pick a winner.
The case for
If AI makes software vulnerabilities easier to find and fix, governments may lose a practical route to access suspects’ devices without weakening encryption for everyone. That could revive pressure for backdoors, making the debate over investigative powers and the security of ordinary users more urgent.
The case against
AI may change how vulnerabilities are found without removing governments’ determination or ability to conduct surveillance; states could develop new techniques or retain other forms of access. On this view, predictions that stronger software will force a return to backdoors are uncertain, and should not be treated as an inevitable consequence of AI-driven security improvements.