Start with the bill itself. Download a PDF from the latest release, if one is attached, or use the Markdown reading copy. Give that document to an AI assistant that accepts it, and ask it to explain the proposal in plain language.
Try these questions:
- “Explain how this bill uses infractions, misdemeanors, and felonies to address misuse of government power.”
- “What is the difference between civil immunity and immunity from criminal prosecution?”
- “Explain the assault and battery definitions in Section 1, with examples.”
- “How does the bill make accountability public?”
- “What are the strongest legal objections to this proposal?”
- “Which exact sections support your answer?”
Ask specific, document-grounded questions — don't just ask an AI assistant to summarize or render a verdict on the bill cold. An assistant answering without the text in front of it is drawing on patterns in its training data about citizen-drafted accountability bills, qualified immunity, and government power generally — and that training data carries its own assumptions, which can run either direction, dismissive or uncritical, independent of what this bill actually says. Give it the document first, ask it to name the exact section behind each answer, and treat any claim it can't point to a section for as unverified.
Compare the answers with the document you supplied. Ask the assistant to distinguish what the bill proposes from what existing law already allows. Check the release version: a PDF snapshot may be older than the working draft, and the source files in Government Accountability and Citizen-Led Investigations may contain changes not yet included in the assembled reading copy.
🚧
Multiple independent studies have measured political skew in mainstream AI chatbots when they're asked for a bare opinion rather than reasoning from a supplied document:
- David Rozado, "The political preferences of LLMs", PLOS ONE (2024) — ran 24 conversational LLMs through 11 political-orientation tests; most scored left-of-center, while the underlying base models (before fine-tuning) did not show the pattern.
- Tavishi Choudhary, "Political Bias in Large Language Models: A Comparative Analysis of ChatGPT-4, Perplexity, Google Gemini, and Claude", IEEE Access (2025) — found ChatGPT-4 and Claude leaning liberal, Perplexity leaning conservative, and Gemini closer to centrist.
- "Language-Dependent Political Bias in AI: A Study of ChatGPT and Gemini", arXiv (2025) — found the direction and strength of the lean shifted depending on the language used to query the model.
- The Washington Post's chatbot investigation, June 2026 — reporters scored six major chatbots on 30 political questions; most answered left-leaning more often than right-leaning, though OpenAI, Anthropic, and Google each disputed or qualified parts of the result.
The studies disagree on which model leans which direction and by how much, and the companies involved dispute some of the figures — but independent researchers keep finding some measurable lean, in both directions depending on the model, the language, and the topic. Treat that as a known property of the tool you're using, not a reason to distrust this bill specifically. It's exactly why the instructions above ask you to hand the assistant the actual text and make it cite sections, instead of asking it for a cold opinion.
📱💰 Use a capable AI model, not just whatever free chatbot is in front of you 💰📱
This document is long and dense — hundreds of pages of interlocking, cross-referenced sections. 🆓
🧾 If getting accurate answers about this bill matters to you, it's worth paying for access to a model capable of reading all of it. Either way: follow the instructions above — hand it the actual text, and make it cite the section behind every claim. 📖✅
Operation Checkmate is a proposed Utah bill that makes defined abuses of government power punishable as infractions, misdemeanors, and felonies. Its purpose is public accountability: officials answer for unlawful conduct under law, before a court, with consequences tied to what they did.
Criminal accountability is needed because internal investigations can be corrupted by the institution’s interest in protecting itself. When colleagues investigate colleagues, supervisors review conduct they approved, and the agency controls the records and public explanation, a process meant to uncover abuse can become a way to conceal it. The institution accused of wrongdoing should not have the final word on whether its officials are held accountable.
The central distinction is simple: qualified immunity and the familiar judicial and prosecutorial immunities are protections against civil liability. Those civil shields do not provide immunity from criminal prosecution. Qualified immunity concerns civil damages; the Supreme Court has also expressly distinguished judges’ and prosecutors’ civil immunity from criminal responsibility for willful violations of constitutional rights. See Harlow v. Fitzgerald and Imbler v. Pachtman.
That distinction is the strategy behind this bill. Define the prohibited conduct, assign an offense level, and require proof of the violation. An official facing a criminal charge cannot turn a civil immunity doctrine into a defense to that charge. The question becomes whether the evidence proves the offense and the required mental state.
The proposal covers a range of conduct because misuse of power comes in different forms. Minor coercion can start at the infraction level. Repetition or a specified aggravating circumstance can make conduct a misdemeanor. Assault, battery, and other serious abuses carry felony classifications under the draft. The precise offense and its elements control; every disagreement with an official does not carry the same charge.
That is the “checkmate”: define misuse of power as an offense, and the ordinary civil immunity shield no longer answers the criminal case. The officeholder must answer the allegation through the legal process, with the same requirement that criminal liability be proved beyond a reasonable doubt.
Criminal accountability for official misconduct already exists in some forms. For example, 18 U.S.C. § 242 criminalizes willful deprivation of federal rights under color of law. Operation Checkmate proposes a broader, detailed state framework naming prohibited conduct, setting levels of responsibility, and assigning consequences.
A legal limit matters: criminalization does not automatically eliminate every form of government immunity. Separate constitutional protections can apply in criminal cases; presidential immunity for official acts is one example recognized in Trump v. United States. A state bill cannot simply override federal constitutional protections. The claim here is that civil immunity does not defeat criminal liability, not that changing an offense’s label resolves every constitutional question.
The draft also proposes changes to civil immunity in Section 5. Those provisions are part of the full proposal. The criminal approach stands on its own distinction: a civil defense does not decide whether someone committed a crime.
The purpose is to end in-house or private accountability as the final answer to conduct that meets the elements of an offense. Misuse of public power should produce an answer the public can examine: what happened, what law applied, what evidence was presented, and what consequence followed.
An internal finding that an employee followed policy does not establish that the conduct was lawful. A supervisor’s approval does not supply authority the law never granted. A private reprimand does not resolve criminal responsibility. Under the proposal, the legal question follows the conduct even when an institution would prefer to describe it as a personnel matter.
A corrupted internal investigation can bury complaints, discount evidence, excuse colleagues, or relabel unlawful conduct as a minor policy issue. Closing an internal investigation must not close the question of criminal responsibility. Defining the conduct as an offense gives it a legal standard that the accused institution cannot settle through its own personnel decision. Prosecution brings that question before a court, where evidence must establish guilt and civil immunity cannot shield the criminal conduct. That is the proposal’s means of holding officials publicly accountable.
Public accountability means an allegation can be evaluated against a stated rule. Evidence can be tested. The accused can answer. A decision can be explained and reviewed. The public has something more concrete than an assurance that an office looked into itself and considers the matter closed.
The same principle applies when several offices participate. If one official initiates an unlawful deprivation, another supplies the paperwork, and another carries it out, the draft examines each person’s conduct and responsibility. Moving the act through a chain of signatures does not, under the proposal, erase the underlying harm or make every participant unreachable.
The bill’s intended result is public legal accountability for misuse of power, with responsibility attached to the actors whose conduct satisfies the law’s requirements. Internal discipline may still serve an employment function, but it cannot substitute for deciding criminal liability. An accusation alone establishes neither guilt nor punishment; the case must still be proved.
The bill uses assault and battery in a specific government context. Both definitions protect rights, liberty, and property. Neither is limited to a physical injury. These are the proposal’s definitions, and should be read as such rather than assumed to describe every existing assault or battery statute.
| Proposed offense | Basic distinction | Draft classification |
|---|---|---|
| Assault | Official conduct puts someone in a position where a reasonable person would believe they may need to defend their rights, liberty, or property. | Class 3 felony |
| Battery | Official conduct actually deprives, restrains, harms, or violates someone’s rights, liberty, or property. | Class 2 felony or higher, depending on severity |
The classifications above are the draft’s stated baselines. The mental-state rules and any applicable offense-specific provisions or enhancements must also be read. The controlling definitions are in Section 1.
Under §1.1.2, assault occurs when a government actor, acting under color of authority, places a citizen in a position where a reasonable person would believe they may be required to defend their rights, liberty, or property. The definition expressly includes having to defend those rights in a court of law.
“Under color of authority” concerns the use or claimed use of official power. The conduct carries the weight of an office, badge, order, or governmental role. An official’s ability to detain, charge, regulate, remove, or otherwise act against someone gives a threat a practical force beyond an ordinary disagreement.
Examples named in the draft include threats of unlawful detention, coercive commands without lawful basis, actions creating imminent fear of a rights deprivation, and abuses of authority that pressure someone to comply beyond lawful bounds. The definition also reaches conduct compelling someone to seek judicial relief or mount a legal defense to preserve their rights.
Consider an official who threatens an unlawful arrest to force someone to surrender a right. The threatened arrest may never happen, but the person has already been put in the position of defending their liberty against official power. That is the kind of conduct the assault definition addresses. Whether a particular incident qualifies depends on the facts and the required mental state.
The court-related language is a substantial part of the proposal. It treats the coercive use of legal proceedings as capable of threatening rights even when no physical confrontation occurs. The reader should evaluate that breadth directly: the draft reaches the pressure to defend rights, including through litigation, before an actual deprivation is completed.
Under §1.1.3, battery occurs when a government actor actually deprives, restrains, harms, or violates a citizen’s rights, liberty, or property. It can happen through direct action or an indirect process that produces the deprivation.
Direct examples include unlawful detention or imprisonment, physical force without lawful justification, enforcement beyond lawful authority, and execution of orders that violate constitutional or statutory rights. The common feature is an actual intrusion on the protected interest.
Indirect battery addresses the same result produced through paperwork or institutional action. The definition names orders, warrants, judgments, holds, referrals, sign-offs, proceedings, and other intermediate means. It also addresses participation in systems or chains of officials that bring about an unlawful deprivation.
For example, an official who knowingly procures an unlawful detention through false paperwork may cause the same loss of liberty as the person who physically carries it out. Under the draft, the fact that the first official worked at a desk does not by itself remove that conduct from the battery definition.
Section 1.1.3.1 develops this principle: a deprivation is not made lawful merely because it passed through an order, administrative finding, automated system, or chain of officials. Responsibility still depends on participation and the applicable mental state. A signature or institutional role is not, by itself, proof of the offense.
The definition also distinguishes unlawful deprivation from lawful protective action. Section 1.1.3.2 requires a factual and legal basis for protective action and describes the care required in taking it. The proposal’s rule is that authority must rest on an actual basis for the action; the possibility that something bad might happen does not supply unlimited power to deprive someone of liberty or property.
The distinction to remember: assault addresses being put in a reasonable position of needing to defend a right; battery addresses the actual deprivation or violation. The proposal applies both concepts to the exercise of government power.
A useful plain-language description of the proposal is citation-level infractions, misdemeanors, and felonies. The draft’s formal term for the lowest tier is infraction. A citation is a way an offense may be charged or noticed, rather than a separate offense classification.
Under §1.1.4, an infraction includes minor coercive conduct, falling short of assault or battery, that would reasonably direct or pressure someone to act contrary to their rights, liberty, or property. The draft includes utterances, gestures, and implied commands without lawful basis.
For example, “move along” is evaluated in its circumstances. The draft’s example concerns a directive issued where the person is acting within their rights and no lawful basis supports the command. The point is the misuse of authority conveyed by the directive, not the words considered in isolation.
The draft classifies that conduct as an infraction, escalating to a Class 1 misdemeanor on repeat occurrence by the same actor or where it forms part of a pattern. Repeated infractions may also be aggregated as assault under the provision’s terms. Other offenses specify their own misdemeanor or felony treatment.
| Level | What the draft uses it to address |
|---|---|
| Infraction | Specified minor coercion and expressly covered negligent conduct. |
| Misdemeanor | Specified repeated, patterned, or aggravated misconduct, including escalation under §1.1.4. |
| Felony | Assault, battery, and other defined serious abuses of official power. |
This is a range of accountability, not a requirement that every official receive a warning and work through each level before a serious charge is possible. Conduct meeting a felony’s elements can be charged at that level. Conversely, the existence of felony provisions does not turn every minor violation into a felony.
The mental-state rules in Section 3 distinguish negligence, recklessness, and intentional conduct. Those distinctions matter because the bill ties liability and punishment to both the act and the actor’s responsibility for it.
- Negligence: failing to exercise the required care and failing to perceive a substantial risk the actor should have perceived.
- Recklessness: disregarding a substantial and unjustifiable risk, ignoring an obvious risk, or proceeding while aware of a possible violation, as defined in the draft.
- Intentional conduct: acting purposefully or with awareness that the result is substantially certain.
Negligence alone is a sufficient, independent basis for liability wherever the draft reaches it. The draft does not require proof of intent or a motive before negligence-based liability attaches. This is the same standard the law already uses for driving under the influence, negligent homicide, and manslaughter — none of which require proof that the actor wanted to cause harm. A failure to exercise reasonable care is enough, and the draft applies that same standard to government conduct.
The default minimum for criminal liability is recklessness, unless a specific offense sets a different standard. Some provisions expressly reach negligence; others require a higher mental state. Intentional and malicious conduct can increase the classification under the draft’s rules.
Section 3.4 requires the prosecution to prove the offense elements and criminal mental state beyond a reasonable doubt. Removing a civil immunity shield does not supply missing evidence or establish guilt. The conduct still has to fit a valid offense, and the prosecution still has to prove its case.
The proposal also rejects internal policy, orders from a superior, and official status as substitutes for lawful authority. An official must be able to identify why an action was permitted. “That is how our office does it” does not resolve whether the conduct deprived someone of a protected right.
This gives the accountability proposal a concrete standard to assess. Readers can ask whether a definition is clear, whether the mental state is appropriate, whether the penalty matches the conduct, and whether the process protects against wrongful conviction. Those questions belong in the review of the actual bill text.
The draft’s definition of government actor reaches the exercise of public authority across roles: police officers, judges, prosecutors, elected officials, regulators, clerks, public administrators, and other covered actors. Section 17 adds offenses tied to particular responsibilities.
That breadth matters because a deprivation can begin in many places. An unlawful detention may involve a street encounter. A loss of property may begin with an administrative order. A coerced surrender of a right may happen in a courtroom, an inspection, or a licensing process. The proposal follows the use of power and its effect on rights.
Its core standard is that public office carries responsibility for the power entrusted to it. The bill’s definitions, proof requirements, and classifications are intended to make that responsibility specific enough to address in a legal proceeding. Its treatment of those details is open to scrutiny and revision as part of this working proposal.
This repository contains proposed legislation, not an enacted law. Read the full text before deciding what to support or change. The README introduces the central argument; the operative sections contain the detailed rules, exceptions, classifications, and remedies.
| Looking for… | Read |
|---|---|
| The assembled proposal (both Acts, combined) | Full bill |
| Definitions of assault, battery, and infraction | Section 1 |
| Mental states and proof requirements | Section 3 |
| Immunity and liability provisions | Section 5 |
| The citizen-enforcement mechanism (Section 9) | Citizen-Led Investigations |
| A section-by-section overview | Advanced overview |
| Answers to common questions | FAQ |
| Current source structure and build instructions | Government Accountability index · Citizen-Led Investigations index |
| Explanations and community guides | Why essays · Community guides |
| How to propose an edit | Contributing |
The bill is two companion Acts — Government Accountability and Citizen-Led Investigations — that reference each other by section number. The root bill_proposal.md is the combined reading copy of both. Source edits in government_accountability/ or Citizen_Led_Investigations/ can be newer. Companion essays may use older section numbers, so check the section indexes when following references. Documentation edits do not require rebuilding the bill.
To sign the petition in support of Operation Checkmate, email citizens.checkmate@gmail.com with subject CHECKMATE and include your first and last name and ZIP code. Your submission counts as an electronic signature supporting the petition. Aggregated counts may be published; your full identity will not be published without separate, explicit consent.
Questions, criticism, and suggested wording are welcome at the same address. You can also open a GitHub issue or propose an edit using the contribution guide. For specific feedback, identify the section and explain what you would change and why.
This is a citizen-written legislative draft. It is not legal advice.