SOURCE: Speech free from government pressure, Origins in the open, Unmasked fakes, Remedies for deliberate harm, Corrections start with government, Evidence you can checkThe Information Trust Act · a citizens’ proposal
Democracy runs on information people can trust. Today, whichever side you’re on, you have reasons to doubt the news, the platforms and the government itself, and bad actors count on it. The SOURCE Act doesn’t tell anyone what is true. It keeps the government’s hands off lawful speech, shows you who is really talking, labels fakes, and makes deliberate lies that hurt people costly, as far as the First Amendment allows.
The government stops leaning on speech, and starts showing its own work.
No official may pressure a platform, broadcaster or bank to bury lawful speech, and every request about content goes on a public log. Agencies cite their sources, correct their own errors, and can’t secretly pay commentators.
You can see who is paying, who is foreign and what is fake.
Content paid for by foreign governments is labeled. Paid political posts name the payer. Bots say they’re bots. Realistic AI fakes of real people must carry labels you can actually see, and big AI services build in ‘made by AI’ credentials.
Deliberate lies that hurt people carry a price. Opinions never do.
Harassment suits over speech are thrown out early. A publisher that refuses to correct a deliberate lie pays the victim’s lawyers. Lying to take people’s money, or about how to vote, is punished. Opinion, satire and honest mistakes are always protected.
Who it covers
The federal government first: every agency and the White House. Private speakers only in the narrow categories courts have always allowed: who is paying, fraud, impersonation and deliberate harm. Outlets, platforms, AI makers and States can choose to join a fuller Information Trust Standard and earn legal protections in return. No agency gets any power to decide what is true.
Federal agencies and the White HouseThe FCCAgents of foreign governmentsPeople paid to post about politicsBot operatorsLarge AI image, voice and video servicesAnyone who knowingly fakes a real personOutlets, platforms and States that choose to join
Why now? Because both sides have felt it.
You may remember different examples depending on where you sit. The fix is the same.
If you lean left, you may remember
Fox paying $787.5 million to settle over false claims about 2020 voting machines.
An FCC chairman’s warning to ABC stations, hours before Jimmy Kimmel was pulled off the air.
Russia’s troll farm reaching as many as 126 million Facebook users.
If you lean right, you may remember
White House pressure on Facebook over COVID posts, and the laptop story buried weeks before an election.
51 former intelligence officials suggesting the laptop was Russian disinformation.
FBI surveillance applications with 17 significant errors, built partly on an unverified dossier.
And both sides remember a fake AI voice of a President telling voters to stay home, and Russian state money secretly funding American creators. The lesson is the same: only 28% of Americans now trust the mass media, 8% of Republicans and 51% of Democrats. Rules that only one side trusts last until the other side wins.
Step 2 · Why you should care
If you can’t trust the source, bad actors decide what you believe.
Manipulators don’t need you to believe their lie. They only need you to stop believing anything, so the loudest or best-funded voice wins. Every other decision a citizen makes, about a vote, a vaccine or a cause, depends on knowing where information comes from. Here is what has gone wrong, from both directions, and what the act does about each.
Forty years of distrust, from both sides
The color on the left edge shows whose worry it mostly was. Each side’s trust fell after a different set of failures, and each blames the other’s sources.
1987Left’s worryThe FCC repeals the Fairness Doctrine, which had required broadcasters to air contrasting views. Talk radio booms; many on the left date today’s partisan media to this vote.
1996Both sidesSection 230 shields online platforms from liability for users’ posts. The same year, Fox News and MSNBC launch.
2004Left’s worryThe New York Times admits its prewar Iraq weapons coverage relied on sources it failed to check.
2004–05Right’s worryCBS airs unauthenticated documents about President Bush’s Guard service weeks before an election; an outside panel faults the network and four people are fired.
2005Left’s worryGAO finds the Education Department’s $240,000 payment to a commentator to promote its policy was illegal covert propaganda.
2016Both sidesRussia’s troll farm reaches up to 126 million Facebook users; a gunman fires a rifle in a D.C. pizzeria while ‘investigating’ the Pizzagate hoax.
2019Right’s worryThe Justice Department inspector general finds 17 significant errors in FBI applications to surveil a Trump campaign aide.
2020Right’s worry51 former intelligence officials say the Hunter Biden laptop story has the earmarks of Russian disinformation; platforms limit the story.
2021Right’s worryThe White House presses Facebook for months to remove COVID posts, including satire.
2022–23Left’s worrySandy Hook families win $1.4 billion over claims the shooting was staged; Fox pays $787.5 million to Dominion over false 2020 claims.
2024Both sidesAn AI clone of President Biden’s voice tells voters to stay home; Russian state media secretly funds U.S. creators; the Supreme Court rejects government coercion in Vullo but dodges the platform case on standing.
2025Both sidesMeta drops fact-checkers; a new order bans federal ‘censorship’; the State Department’s disinformation office and the FBI’s foreign-influence task force are shut down; CBS’s owner pays $16 million to settle a President’s suit; Pentagon reporters walk out over access rules.
2026Both sidesA party committee runs an AI video of a Senate candidate with a faint disclosure; days before the midterms, Russia-linked fakes target Senate candidates.
Step 3 · What it does
Six promises, one for each letter.
The act has 10 packages in two parts: six that are binding law (the Trust Floor) and four that outlets, platforms, AI and device makers, and States choose to join (the Information Trust Standard). Together they keep six promises. Tap a number to read the package behind it.
What is binding law, and what each chooses
The First Amendment rightly limits what Congress can require of private speakers. So the binding part of the act mostly binds the government itself, and reaches private speakers only where courts have always allowed: who is paying, fraud, impersonation and deliberate harm. Everything about how news is labeled or feeds are run is a standard each outlet or platform chooses.
The Trust Floor: binding law
No federal pressure on platforms, broadcasters or banks over lawful speech
Every government request about content on a public log; the speaker can sue
Agencies cite sources, correct errors, and never pay for covert promotion
Foreign-government content labeled; covert foreign operations a crime
Paid political posts name the payer (over $1,000 a year); bots disclose
Knowing fakes of real people must carry visible labels; big AI services add credentials
Anti-SLAPP; refuse to correct a deliberate lie and pay the lawyers; fraud and voting lies punished
No agency decides what is true; Section 230 unchanged
The Information Trust Standard: each chooses
Outlets and creators: label news, analysis and opinion, show sources, correct fast
In return: no presumed or punitive libel damages after a prompt correction
Platforms: user-chosen feeds, an ad library, researcher access, human appeals, context notes
In return: one national rulebook instead of fifty, enforced by State attorneys general too
AI and device makers: credentials on AI output and on camera originals
In return: no liability for users’ fakes, one national labeling rule
States: grants to teach source-checking skills, never verdicts
Why they would join: each package offers protection that is worth real money, from lower libel exposure to a single national rule. Why they can’t cherry-pick: the protection comes only with every practice in the package, and the government never judges content, only whether the practices are followed.
No agency, board or contractor gets power over what anyone says. Every finding that a statement is unlawful is made by a court, under the narrow sections of the act, after a fair hearing.
What it does not do
It doesn’t let any government agency decide what is true, or rate news outlets.
It doesn’t punish any opinion, satire or honest mistake.
It doesn’t bring back the Fairness Doctrine or tell anyone to present ‘both sides’.
It doesn’t repeal or rewrite Section 230, or force platforms to carry or remove lawful posts.
It doesn’t require any outlet to label its news or opinion; that is a standard outlets choose.
It doesn’t require donor lists; disclosure means naming who paid.
It doesn’t apply to the November 3, 2026 election.
It doesn’t need a constitutional amendment.
Already in the companion acts
SOURCE is a companion to the Public Trust Act, whose Package 13 shows who pays for political ads, Package 15 protects statistics and science from political pressure, Package 19 bars government retaliation for speech, journalism or research, and Package 2 covers payments to sitting officials. For the CLEAR Act, SOURCE now holds a single set of rules on government pressure and AI fakes, for all lawful speech, all year, while CLEAR keeps its election-season extras: faster publication of requests and quicker court orders against fakes of candidates, and warnings to election officials. SOURCE doesn’t repeat the rest.
Step 4 · Where do you stand?
Check the reforms you support.
Your choices stay in your browser. Use them in step 6 to write to your representatives.
Part 1 · The Trust Floor: binding law
Binding on the federal government, and on private speakers only where courts have always allowed. Each applies the same way whichever party is in power.
Part 2 · The Information Trust Standard: each chooses
These apply only to outlets, platforms, makers and States that join. Checking one means you’d want them offered and would like to see your outlets, platforms and State join.
You support 0 of 15 reforms.
Step 5 · What you’ll hear
The pitches, the pushback and who’s paying.
You’ll hear this proposal praised and attacked, sometimes by people who haven’t read it. Here is how each side is likely to talk about it, the strongest honest criticisms, and the answers.
How each side can make the case
The pitch to voters on the left
“No FCC chairman gets to threaten a network’s licenses over a comedian, or any program.”
“Foreign governments and bot farms have to show their faces.”
“The people behind the Sandy Hook and Dominion lies would have paid sooner, and liars who won’t correct pay the lawyers.”
“AI fakes of real people must carry labels you can actually see, even in campaign ads.”
The pitch to voters on the right
“No more White House pressure on Facebook to bury stories, and every request is public.”
“No government ‘disinformation’ boards and no tax-funded blacklists of conservative outlets.”
“When officials call a story ‘Russian disinformation’, they have to show their evidence.”
“Harassment suits against speakers get thrown out, and nobody is prosecuted for a joke.”
Fair criticisms, and the answers
Each criticism is stated in its strongest honest form. This proposal was tested against every one of them before publication.
How money and media are likely to play it
Expect different messengers to cheer different packages and attack others. Knowing who stands to gain or lose helps you read the coverage.
Messenger
Likely to support
Likely to attack
What to watch for
Right-leaning cable, talk radio and websites
The ban on government pressure, the public log, ‘show your basis’ for disinformation claims, no blacklists
Paid-post disclosure, the Foreign Influence Center
Calling the whole act a ‘censorship bill’ because of its name
Left-leaning cable, podcasts and websites
Broadcast neutrality, foreign-content labels, deepfake liability, penalties for deliberate lies
Leaving Section 230 alone; limits on officials contacting platforms
Calling the act ‘toothless’ for not policing misinformation
Independent podcasters and creators
No pressure on platforms, fair process, user-chosen feeds
Paid-post disclosure (for sponsored political content)
Clips that skip the $1,000 threshold and the ‘payer only’ rule
Large platforms
One national rule, liability protection
Researcher access, ad libraries, appeals deadlines
‘Security’ objections that are really about secrecy
AI companies
A single national labeling rule; safe harbor for members
Mandatory credentials for the largest services
Claims that provenance is impossible; open standards already exist
Short, personal messages that name specific reforms get noticed. Your message below uses the reforms you checked in step 4.
Share it
Send the page, or copy a one-page summary to paste into an email, a post or a document you can print.
Step 7 · The details
The 10 packages, in two parts.
Part 1 is the Trust Floor, which is binding law. Part 2 is the Information Trust Standard, which outlets, platforms, makers and States choose to join. Open any package to see what it does, why, what critics say, and the bill text itself.
Related efforts
Who else is working on this.
The SOURCE Act builds on decades of work by free-speech lawyers, press-freedom groups, technologists and lawmakers in both parties. Most of its ideas aren’t new. What’s new is putting each side’s fix in the same bill: no government pressure on speech, which the right has demanded, together with real consequences for deliberate lies and foreign operations, which the left has demanded.
How this work fits with theirs
It borrows from both camps. The anti-coercion rule codifies a unanimous 2024 Supreme Court decision; researcher access and ad libraries come from a bipartisan Senate bill; content credentials come from an open industry standard; the voting-lies section follows the path courts have said is open.
It complements, rather than competes. Most groups below defend one value: free speech, a free press, privacy or information integrity. SOURCE tries to protect all four at once, and leaves campaign-ad disclosure and anti-retaliation to the Public Trust Act.
What SOURCE adds
Each side’s fix in the same bill. A ban on government pressure comes with labels on foreign propaganda; anti-SLAPP protection comes with fees for liars who refuse to correct.
Rules for the government first. A public log, sourced statements, corrections, and proof behind ‘disinformation’ claims.
No referee of truth. No agency gets power over content; courts decide the narrow cases.
Incentives instead of mandates for outlets, platforms and AI makers, through an opt-in standard.
Draft text, sources and a plan to pass it, from an independent citizen with no party, PAC or donor directing it.
Listed for information only. This project is not affiliated with, endorsed by or coordinating with any of these organizations or lawmakers, and none has reviewed this proposal. Descriptions are based on each group’s public materials.
Active: changes in the last 1–2 months · Quiet: last change 3–6 months ago · Dormant: nothing new in 6+ months. Affiliation follows how each group describes itself, or the parties of a bill’s sponsors; courts, agencies and States are unaffiliated. A ring slightly off center marks a nonpartisan group that outside critics see as leaning one way; a gold ring marks a libertarian group.
Creates a federal right over AI replicas of a person’s voice and likeness, with bipartisan sponsors in both chambers. SOURCE complements it: NO FAKES covers commercial replicas; SOURCE covers deception about real people and labels.
Bipartisan bill for vetted researcher access to platform data and public ad libraries. SOURCE offers the same in its Open Feeds standard, paired with a national rule.
Federal reporter’s shield law; passed the House unanimously in 2024. Listed in the Public Trust Act’s media options; SOURCE adds anti-SLAPP and broadcast neutrality.
Executive Order 14149 bars federal ‘censorship’; a December 2025 order directs challenges to State AI laws. Orders can be reversed by the next President; SOURCE puts the speech rules in statute and offers platforms and AI makers a national rule only in exchange for user protections.
The Supreme Court held that officials may not coerce private companies to punish speech, but threw out the platform case on standing. Section 103 fixes the standing problem.
Platforms’ curation is protected expression; requiring them to report how they handle State-defined categories is compelled speech. That is why the platform rules are an opt-in standard.
California’s election deepfake laws struck down; a ‘vote by text’ meme conviction reversed. Sections 401 and 505 are drawn to the narrow paths the courts left open.
Know an organization or bill that belongs here? Suggest it through the feedback form.
About this page
About & sources
Created by a deeply concerned citizen with the help of generative AI tools.
This page is the public summary of a citizens’ proposal called the SOURCE Act, formally the Information Trust Act. It has not been introduced in Congress. The draft legislative text still needs review by First Amendment and media lawyers on the left and right, by press-freedom and civil-liberties groups, by technologists, and cost estimates from the Congressional Budget Office.
It is a companion to the Public Trust Act, which sets the same rules for every official, the STEADY Act on money and prices, the LEVEL Act on the courts, and the CLEAR Act on elections. All five work under today’s Constitution.
It was published days before the November 3, 2026 midterm elections. Nothing here is about who should win them, and none of it would apply to them.
Content matches the plan draft v0.2 of November 1, 2026. Facts are sourced below; the type of each source is noted so you can judge it yourself.
An open letter to members of Congress
To every Senator and Representative, of every party:
Your constituents no longer trust what they read, watch or hear, and each side can name the moments that taught it to doubt: a government that leaned on platforms, a network that aired false claims, intelligence officials who vouched for a theory, a foreign troll farm, a fake voice of a President. Bad actors at home and abroad count on that doubt. When citizens can’t tell sourced fact from opinion or fabrication, the loudest and best-funded voice wins.
The SOURCE Act asks you to fix the rules, not to win the argument. Keep the government’s hands off lawful speech, and put every request to a platform on the record. Make the government show its sources and correct its errors. Show who is really talking: foreign governments, paid promoters and bots. Label fakes of real people where people can see them. Make deliberate, harmful lies costly in the narrow ways the Constitution allows, and never let any agency decide what is true.
Each package can pass on its own. We ask you to co-sponsor them with a member of the other party, to keep supporting them when your party holds power, and to tell us publicly where you stand on each one.
Respectfully, Your constituents
Sources
Privacy statement
This page sets no cookies and runs no analytics or ads. Your checklist choices and message draft stay in your own browser’s storage and are never sent anywhere. “Give feedback” opens a Google Form, and “Become a sponsor” opens Ko-fi; each has its own privacy policy.
Sponsorship terms
Sponsorships support this independent citizens’ project: hosting, research and printing. They are gifts: final, non-refundable and not tax-deductible. Nothing is given in return, and sponsors have no say over the content. Spending is reported each year by category.