‘All men are created equal’: America has lost its values. It’s time to go back to the founding text | Ted Widmer - The Guardian

The Declaration of Independence isn't just a political document - it's the original open-source license for democracy. And we've silently forked our values into a proprietary, buggy branch. As America approaches the 250th anniversary of its founding, critical voices like historian Ted Widmer remind us in "'All men are created equal': America has lost its values. It's time to go back to the founding text" that the core promise of equality has eroded. But this isn't just a political crisis - it's a technology crisis. From algorithmic bias to digital surveillance, the systems we've built often run on code that contradicts the very principles the founders etched into parchment. If we're serious about reclaiming those values, we need to audit our infrastructure the way we'd audit a critical deployment. Let's trace the fault lines.

When the source code of democracy was written

On July 4, 1776, a group of engineers of political thought - Jefferson, Adams, Franklin - released the first commit of a nation built on self-evident truths. The phrase "all men are created equal" wasn't just rhetorical garnish; it was a design constraint. Like a framework's core principles, it set boundaries for every subsequent feature: representation, due process, freedom of speech. Modern software teams know the pain of abandoning a well-architected foundation. We've all inherited that legacy codebase where early assumptions were sound but later patches introduced inconsistency. America is that codebase. The 250-year-old commit log shows a project that began with clean abstractions - inalienable rights, social contract - and then accumulated technical debt through slavery, disenfranchisement. And corporate capture.

What Widmer's Guardian piece captures is the realization that the founding text itself functions like a README md for a nation. It declares the purpose, the dependencies (the consent of the governed). And the exit strategy (the right to alter or abolish a destructive government). When we deviate from that spec, the system behaves unexpectedly. Today's symptoms include algorithmic suppression of minority voices, mass surveillance chilling dissent. And monopolistic platforms extracting value without accountability. The bug isn't in the original source - it's in the runtime environment we've allowed to grow around it.

Algorithmic inequality: When code betrays the creed

If "all men are created equal" were a software requirement, modern AI systems would fail the test. In production environments, we've seen hiring algorithms that discriminate against women (Amazon's 2018 resume screener), facial recognition that misidentifies Black individuals at higher rates (MIT Media Lab's 2019 Gender Shades study), and predictive policing tools that disproportionately target minority neighborhoods (ProPublica's investigation on COMPAS). These aren't edge cases - they're systemic failures arising from training data that encodes historical inequality and loss functions that improve efficiency over fairness.

Consider the technical parallel: the Declaration asserts a universal equality as an axiom. In contrast, most machine learning models assume a utility function that can accept trade-offs between accuracy across demographic groups unless explicitly constrained. When we deploy such models in criminal justice, housing. Or credit, we're implicitly editing the founding text to say "all men are created equal, except where it's cheaper not to be. " Returning to the founding values means adopting fairness constraints as first-class citizens in our model development lifecycle - akin to how the Constitution's Bill of Rights acts as an immutable layer on top of statutory law.

The surveillance state and the Fourth Amendment: A regression from 1776

Jefferson's generation understood that privacy wasn't a luxury - it was a prerequisite for liberty. The Fourth Amendment's protection against unreasonable searches and seizures was a direct response to the British crown's writs of assistance (essentially, general warrants). Fast-forward to 2025. And we have a surveillance architecture that would make King George III envious. The NSA's Section 702 program, authorized under FISA, vacuums up millions of communications from non-U. S persons every year. And the government regularly obtains warrantless access to location data through third-party doctrine arguments.

From a software engineering perspective, what we've built is a globally distributed monitoring system with insufficient access controls and no proper audit trail for citizens to verify compliance. The founding text's commitment to "security of the people" against government overreach has been replaced by a threat model that treats the state as benign and the citizen as a potentially hostile actor. To restore the original values, we need to apply zero-trust architecture to governance - assume that any agency with surveillance powers will eventually abuse them unless constrained by cryptographic safeguards - judicial oversight. And public transparency. Projects like the EFF's surveillance self-defense guides show that the technical means for privacy exist; what's missing is the political will to mandate them.

Open source vs. closed platforms: The battle for equal access

The Declaration's vision of equality also has a profound analog in software distribution. The founders believed knowledge and power should be decentralized - hence the Federalist Papers published in newspapers, not held in a single royal archive. Today, we've replaced the printing press with platform monopolies that control the discourse. Apple's 30% tax on in-app purchases, Google's search algorithm shaping election results, Facebook's news feed determining which stories go viral - these are de facto edicts issued by private sovereigns with no democratic accountability.

Open-source software directly embodies the founding ethos: the source is available for everyone to inspect, fork. And improve. When we move from a world where "all code is created equal" to one dominated by closed platforms with opaque ranking systems, we repeat the error of replacing a representative republic with an algorithmic aristocracy. The remedy is to mandate interoperability and data portability for large platforms - a principle akin to the Commerce Clause that ensures no single entity can choke innovation. The W3C Decentralized Identifier framework shows that we have the standards; we need the regulatory enforcement to make equal access a feature, not a bug.

Returning to the source code of American values

Widmer's call to go back to the founding text is fundamentally a call to refactor. In software, when a codebase has drifted so far from its original architecture that patches break more than they fix, we don't abandon the project - we revisit the core abstractions and re-implement them with modern best practices. For America, that means auditing every major institution - from the Supreme Court to the surveillance apparatus - against the Declaration's equality axiom. Does the algorithm distributing SNAP benefits treat all applicants equally? Does the predictive model in child welfare services systematically flag families of color, and these are engineering questions with ethical weight

We can operationalize this by adopting constitutional impact assessments for any automated decision system used by government. Similar to how data protection impact assessments are required under GDPR, a constitutional AI audit would verify that a model's output doesn't violate equal protection. The technical community has already developed frameworks like Fairlearn and AI Fairness 360 - we just need to mandate their use in federal procurement. That's a concrete step toward making the founding text executable again.

Lessons from the Second Continental Congress for today's tech governance

The body that adopted the Declaration - the Second Continental Congress - was a hastily assembled group of delegates with competing interests who but managed to produce a coherent, durable document. NPR's recent piece on whether Congress is living up to that legacy highlights a key insight: the founders didn't wait for perfect consensus; they debated fiercely and then committed to a framework that allowed future amendments. Modern technology governance suffers from the opposite problem: paralysis in the face of complexity, preferring voluntary self-regulation over binding rules.

In engineering, we know that a system without explicit constraints will drift toward whatever behavior is easiest to add, not what is most just. Today's AI governance landscape is a collection of voluntary frameworks - the White House's AI Bill of Rights, the EU's AI Act (still in implementation) - that lack the enforcement teeth of the original Constitution. The Second Continental Congress understood that writing down rights isn't enough; you need checks and balances (the Supreme Court, the electoral process) to enforce them. We need a similar separation of powers in tech: an independent algorithmic accountability office that can audit models, a dedicated court for digital rights disputes and a transparent registry of government-used algorithms.

FAQ: The founding text and modern technology

• What does "all men are created equal" mean for AI ethics? It means that automated systems should be designed to avoid disparate impact across demographic groups. In practice, this requires fairness metrics, diverse training data. And regular bias audits. The principle is a non-negotiable constraint, not an optimization objective.
• Can an algorithm violate the Declaration of Independence? Yes, if its outputs systematically deny equal protection of the laws. For example, a predictive policing model that directs more patrols to Black neighborhoods based on historical arrest data is effectively re-establishing a regime of unequal treatment that the Declaration condemns.
• How can open source help restore founding values? Open source enables transparency, auditability. And decentralized control - all of which align with the Founders' suspicion of concentrated power. Licenses like the GNU GPL even include a "share alike" clause that mirrors the idea that improvements should benefit everyone equally.
• What's the biggest tech-related threat to America's founding values today? The combination of mass surveillance and algorithmic gatekeeping. When the government can monitor all communications and platforms control what speech reaches the public, the rights to free expression and due process become theoretical.
• Is it realistic to apply 18th-century ideas to 21st-century AI? The principles are timeless, but the implementation must evolve. Just as the First Amendment has been applied to the internet, the equality clause of the Declaration can guide how we regulate algorithms. The technical tools exist - what we lack is the collective commitment to enforce them.

Conclusion: The refactoring we need

The Guardian piece by Ted Widmer - "'All men are created equal': America has lost its values. It's time to go back to the founding text | Ted Widmer - The Guardian" - is more than a historical reflection. It's a debugging report for a system that has strayed from its spec. The good news: we have the original source code. The better news: we have the tools and the talent to refactor. Engineers, data scientists. And product managers can lead this effort by demanding fairness, transparency. And accountability in every project we build. It's time to stop treating the Declaration as a museum artifact and start treating it as a living architecture document. Read it, diff it against today's reality, and commit the changes.

If we don't, the next 250 years will be governed by algorithms that were trained on the very inequalities the founders hoped to abolish. That's a future none of us should deploy,

What do you think

If the Declaration were a software license, would you accept it for the AI systems you're building today,? Or would you need a copyleft clause for equality?

Should a constitutional AI audit be a mandatory step in federal procurement,? Or would that stifle innovation and slow down progress unfairly?

The founders debated the balance between liberty and security for months in 1776, and given modern surveillance capabilities,Where should that line be redrawn - and who should draw it?