Schneier on Security

At the Infosec World 2025 conference last week, AI dominated discussions and vendor displays. One sparsely attended speaker joked that including AI in the title of his talk would have drawn a larger crowd.

When I heard about Rewiring Democracy: How AI Will Transform Our Politics, Government, and Citizenship (MIT Press) by Bruce Schneier and Dr. Nathan Sanders, I expected a harsh critique of AI’s impact on democracy, but the book instead presents a nuanced thesis on how AI will transform, rather than simply threaten, our political systems.

Schneier needs no introduction to those in the world of Information Security. His co-author, Sanders, is a data scientist, physical scientist, and a teacher at the Berkman Klein Center at Harvard University.

To me, the authors take an unexpectedly optimistic view of AI’s potential impact on democracy. Their main thesis is that, although AI can be used by oppressive regimes to cause harm, it can also be deliberately leveraged to support democratic values and processes. They provide an insightful exploration of how AI, when properly governed, could bring positive transformation to democracy.

Although AI is still in its infancy, it has already altered much of the democratic process. The authors believe it can be harnessed to strengthen, rather than undermine, democracy.

While governments often work quite slowly and are immune to change (consider the amount of code written on COBOL and FORTRAN running on government systems), the authors write that many governments are already actively using AI. This includes those involved in drafting legislation, judges using it in legal rulings, and more.

The authors write that AI can help make laws more transparent and more consistent. With its superhuman attention spans, AI excels at enforcing syntactic and grammatical rules. It can be effective at drafting text in precise, proper legislative language or at offering detailed feedback to human drafters.

Small details are important and can have significant legal implications. Just recently, a seemingly slam-dunk case in the UK where researchers were alleged to have passed politically sensitive information to a Chinese intelligence agent, which was then provided to a senior member of the Chinese Communist Party.

The head of the Crown Prosecution Service (CPS) said the case collapsed because evidence could not be obtained from the government, referring to China as a national security threat. CPS said that, while there was sufficient evidence when charges were initially brought against the two men in April 2024, a precedent set by another spying case earlier this year meant China would need to have been labelled a “threat to national security” at the time of the alleged offences.

That is just one example of many in which AI can target minutiae in a law and avoid massive consequences later.

It’s not until chapter 27 that they write that security is the most significant major barrier to using AI in democratic applications, and that is something no one seems to be talking about. There’s no trust without security, and the authors write that security appears to be an afterthought in today’s AI development.

That was an issue in 2002 when Bill Gates sent his celebrated email to all Microsoft employees announcing the creation of the Trustworthy Computing (TwC) initiative.TwC was created to put security and trust at the forefront for all Microsoft employees. While Microsoft started Trustworthy AII last year, the industry is still playing catch-up when it comes to securing AI. 

The final part of the book deals with ways to ensure that AI benefits democracy. The authors write that technology will not solve democracy’s problems, but it already has a powerful influence over democracy. They want to shape the emerging technology of AI into a force that can promote good democratic governance, to be used with the best interests of democracy in mind.

They conclude that to build AI that benefits democracy and brings more positive outcomes than harm, society needs organizing principles for those developing and operating AI. These are the same principles any publicly accountable institution must embrace: being capable, available, transparent, responsive, debiased, secure, and non-exploitative.

Democracy has always been inextricably linked with technology. The authors conclude that whether we like it or not, AI is destined to bring changes to our political life and systems of governance.

When it comes to AI, the hype and misunderstanding are overwhelming. The authors cut through all that and provide the reader with a pragmatic, honest look at how AI has impacted and will continue to impact democracy and the democratic process. If you want to understand AI’s true power to strengthen and empower democracy, this is a fascinating and essential read.