TL;DR
Bruce Schneier and Nathan Sanders argue that AI is already embedded across the executive, judicial, and legislative branches, shaping decisions from health-care approvals to legal interpretation. They warn that how policymakers adopt AI — to decentralize power or to concentrate it — will determine its democratic impact.
What happened
In an essay published December 23, 2025, security technologist Bruce Schneier and data scientist Nathan Sanders outline how artificial intelligence has quietly become part of government practice across three branches. In the executive sphere, private insurers and Medicare Advantage operators are using algorithms and modern AI tools to review and sometimes deny medical services, with federal guidance and recent administrative moves loosening prior restrictions and creating financial incentives tied to AI-enabled denials. In the judiciary, judges in multiple countries have begun consulting AI to interpret statutory language, with documented instances in Colombia (2023), a U.S. federal circuit judge (2024), and a D.C. appellate court (2025); judges may consult these tools without being required to disclose their use. In legislatures, AI has been used to draft laws and assist parliaments, including a law reportedly written entirely by AI in Brazil (2023) and bespoke models for parliamentary work in France, while many U.S. state staffers now use or consider using AI. The authors conclude that AI amplifies power — for beneficial or harmful ends — and that the pressing challenge is social and political rather than purely technical.
Why it matters
- AI already affects high-stakes administrative decisions, including health-care approvals that can have life-or-death consequences.
- Judicial use of AI to interpret law raises transparency and accountability questions when use is not disclosed.
- Legislative adoption of AI can either concentrate power in leadership or expand capacity for individual representatives and civic input.
- Policy choices now — about incentives, disclosure, and governance — will shape whether AI strengthens or weakens democratic institutions.
Key facts
- Essay authors: Bruce Schneier and Nathan Sanders.
- Publication date noted in the source: December 23, 2025.
- Private insurers and Medicare Advantage operators are using algorithms and modern AI tools to review, approve, and deny medical services.
- Biden-era CMS guidance largely permitted AI use by Medicare Advantage operators; subsequent administrative actions removed some guardrails and introduced financial incentives tied to AI-enabled denials.
- The Trump administration issued an executive order limiting state abilities to set consumer and patient protections around AI, according to the essay.
- Judicial examples: a Colombian judge used AI in 2023; U.S. Circuit Judge Kevin Newsom used AI in 2024; a D.C. Court of Appeals used ChatGPT in 2025, per the authors.
- Judges may consult AI tools while drafting opinions and might not be obliged to disclose such consultation.
- Legislative examples include the first law said to be written entirely by AI in Brazil (2023) and a French government model built to assist Parliament.
- By the end of the year referenced in the essay, roughly 20% of U.S. state-level staffers reported using AI and about 40% were considering it.
- Civic-technology examples cited: Make.org (European consultations), Scotland’s Comhairle project, and Japanese Diet member Takahiro Anno with Team Mirai.
What to watch next
- Whether federal and state policies will require disclosure when judges or courts consult AI tools.
- How CMS and other agencies adjust incentives and guardrails that shape insurer use of AI in medical authorizations.
- Whether legislative offices adopt AI in ways that decentralize policymaking and increase constituent participation or that centralize influence in party leadership and private interests.
- not confirmed in the source
Quick glossary
- Large Language Model (LLM): A class of AI systems trained on large text datasets to generate or analyze human-like text; often used for drafting, summarizing, or answering questions.
- Medicare Advantage: A type of Medicare health plan offered by private companies that contract with Medicare to provide Part A and Part B benefits.
- Prior authorization: An approval process used by insurers that requires providers to obtain permission before certain medical services will be covered.
- Civic technology: Digital tools and platforms designed to support public participation, government services, or deliberation in policymaking.
Reader FAQ
Are we already being governed by AI?
The authors say we are already being governed with AI in many parts of government and that more adoption is likely.
Is AI fully running government decisions now?
No — the essay says we are not going to be fully governed by AI anytime soon.
Do judges have to disclose when they use AI?
The essay notes a judge may be under no obligation to disclose consulting AI while drafting opinions.
Has any law been written entirely by AI?
According to the authors, a law written entirely by AI was passed in Brazil in 2023.
ESSAYS DEC 23, 2025 5 MINUTES Are We Ready to Be Governed by Artificial Intelligence? By Bruce Schneier, Nathan Sanders Experts Bruce Schneier and Nathan Sanders explore how Artificial Intelligence…
Sources
- Are We Ready to Be Governed by Artificial Intelligence?
- Are We Ready for Artificial General Intelligence?
- Why AI Hive Minds Will Be Needed To Attain AGI
Related posts
- ACCA to end most remote exams from March, citing AI-driven cheating surge
- How California built one of the world’s largest public-sector IT systems
- The New Surveillance State Is You: Citizens Now Turn Tech on Police