Bryan Jacoutot: Bitcoin's Timestamp Is the Last Proof of Truth

↓ Jump to the video and timestamps

I've been making AI-generated video at TFTC for over a year now, and watching the models improve in real time. The progress in 14 months is astonishing. If you project that curve forward even conservatively, we are heading toward a world where you literally cannot tell whether a video is real or fabricated just by looking at it. We are almost there.

That's the context in which I brought Bryan Jacoutot on. Bryan is a commercial litigator at Clark Hill who's been deep in the Bitcoin space for close to a decade, and he's been working with Carlos Cordon's team at Simple Proof to do something I think is wildly underappreciated right now: hash digital artifacts into Bitcoin's timechain to prove they existed in a specific state at a specific moment. He and Rafael Cordon have taken that work all the way to the federal courts, filing an 11-page comment proposing a "rebuttable presumption of authenticity" standard for Bitcoin-timestamped evidence in the Federal Rules of Evidence. That's the conversation. If you care about election integrity, courtroom evidence, or just the basic ability to know what's real, you want to hear this one.

Key takeaways

• AI video is already nearly indistinguishable from real footage, and the gap closes every month. I've watched it happen firsthand at TFTC. The window to build infrastructure of digital truth is open now, and it will not stay open forever.
• Bitcoin's proof-of-work is already doing this job for free. The energy and computation securing Bitcoin transactions is the same infrastructure you're piggybacking on when you hash a document into the chain. No additional cost, no additional energy.
• Screven County, Georgia is the working proof-of-concept. Election Supervisor Stacy Scott hashed certified election results and drop box surveillance video into Bitcoin's blockchain in real time, with roughly a 1-2 second delay. That county now has presumptively valid evidence that others don't.
• Bryan and Rafael Cordon filed an 11-page comment to the Federal Rules of Evidence Advisory Committee proposing a "rebuttable presumption of authenticity" for Bitcoin-timestamped digital evidence, a change that could eliminate the bulk of the discovery costs generated by disputed digital documents.
• This has to be built before the dispute, not after. If you start timestamping because someone challenged your evidence, you've already lost. The whole point is making it a standard business practice now.
• The Library of Congress, AI training datasets, and every contract ever signed are all candidates for this treatment. The question is whether we hash the real record of civilization before generative AI makes it impossible to know what was original.

Why You Can't Trust Video Anymore, And Why That's Bitcoin's Problem to Solve

I want to be clear that I'm not speculating about the future of deepfakes. I'm describing what I've already seen. We've been producing AI-generated video content at TFTC for over a year, and the models we're working with today are dramatically better than what existed 14 months ago. The pace of improvement is not linear.

You project that forward and the destination is a world where no visual inspection can tell you whether footage is real.

Bryan came at this from a different angle, but he's arrived at the same place. His framing is that we're already seeing what he calls "truth decay", and the danger runs in both directions. People are starting to believe AI-generated videos are real, but they're also starting to disbelieve real videos. Both failure modes are catastrophic.

Once the population loses the ability to trust what it sees, an adversarial actor doesn't even need to produce a convincing fake. The confusion itself becomes the weapon.

His framing on that was blunt: if you want to topple civilization, you get people to question everything around them. Get them to not believe what they see, and then get them to believe things that are fake. Nation-states and bad actors don't need a perfect deepfake. They need enough doubt to paralyze the public square.

That's the problem Bitcoin's timechain is positioned to solve, not by making fake video impossible, but by making it possible to prove what the real video looked like before anyone touched it.

How Open Timestamps Actually Works (Without the Math PhD)

The protocol Bryan works with is built on Peter Todd's Open Timestamps Protocol, introduced around 2016. The mechanism is straightforward.

Take any digital file, a video, a contract, a PDF of certified election results. Run it through SHA-256, the same cryptographic hashing function that underpins Bitcoin itself. That produces a fixed-length hexadecimal hash, a string that looks a lot like a Bitcoin address. That hash gets appended to a Bitcoin transaction and written into a block.

Once it's in the chain, two things are simultaneously verified: the content of the file that produced that hash, and the block it landed in, which anchors it to a specific point in time.

If anyone alters the file afterward, even by a single character, the hash won't match. You can go back to the chain, compare, and know immediately that something changed. The comparison is instant and irrefutable.

The data footprint is tiny. Any file, regardless of size, reduces to the same small hash. Bryan made this point explicitly: you're not writing the document into Bitcoin's blockchain, you're writing a fingerprint of it.

That's a meaningful distinction for anyone worried about the ongoing debate over what data belongs in Bitcoin blocks. This protocol stores only a fingerprint of the file in the chain, not the file itself.

The computing power that secures those blocks, the proof-of-work that makes Bitcoin's ledger immutable, is what makes the timestamp trustworthy. You're piggybacking on infrastructure that's already running, already battle-tested, at essentially no additional cost.

Guatemala, Screven County, and the Proof That This Actually Works

Bryan's entry point into this work was Carlos Cordon's team at Simple Proof and what they did in Guatemala. The setup: election records get reduced from physical to digital form, and once they're digital, Simple Proof hashes them immediately. Any alteration that happens after that point is detectable by comparing hashes. Jacoutot's account is that the hashing caught the incumbent attempting to manipulate records after the fact. The altered records contradicted the hash.

That's the demonstration. Once you hash, you have a fixed anchor. Anything downstream of that anchor can be verified or falsified against it.

Screven County, Georgia is the American application. Stacy Scott, the county's election supervisor, hashed the certified 2024 election results into Bitcoin's blockchain at the moment of certification. That gave her something other county supervisors don't have: a cryptographic anchor that freezes the official record in time, immune to after-the-fact alteration.

But the more interesting work is the drop box surveillance video. Georgia requires drop boxes to be monitored when operational, and a state election board rule allows livestreaming that surveillance footage. That rule is well-intentioned, transparency, auditability, show people nothing is happening.

The problem is that a publicly accessible livestream is also a raw material for a generative AI edit. Capture the footage, run it through a model, make it look like ballot stuffing or voter intimidation is happening, and you've got a convincing piece of disinformation with no obvious seam.

What Screven County did: hash the surveillance video into Bitcoin in real time, at roughly a 1-2 second delay from capture. Under Georgia's Open Records Act, that video is public once it's finalized, and anyone can access it. But now there's a Bitcoin hash of what the video actually looked like at capture.

When someone presents a manipulated version, you compare hashes. No match, no credibility. Case closed in minutes, not months.

Bryan made the point that Screven County is a relatively rural county without a massive IT department. That's intentional. If you can implement this there, you can implement it anywhere.

The response from other Georgia officials so far: universally interested, consistently too busy. Bryan's read is that we need to get this statewide before an incident forces the conversation, because once you need it reactively, you've already lost ground. The Guatemala case proved that. The whole value of the protocol is that it's there before anyone tries to dispute the record.

Why This Needs to Be in the Federal Rules of Evidence

Bryan and Rafael Cordon filed an 11-page comment to the Federal Rules of Evidence Advisory Committee proposing a specific standard: a rebuttable presumption of authenticity for digital evidence that's been timestamped into Bitcoin's blockchain.

Here's what that changes. Under the current framework, if a document gets challenged as fraudulent or altered during litigation, the process looks roughly like this:

1. Opposing party challenges the document's authenticity.
2. Judge requests briefs on the relevant law.
3. Targeted discovery is ordered to probe the specific document.
4. Experts are retained to assess the evidence.
5. Hearing is held, possibly followed by an appeal of the ruling.

Bryan walks through this from his actual practice, he litigates commercially, and the discovery process is where the money gets burned. Months of time. Hundreds of thousands of dollars. Hourly attorneys, retained experts, potential appellate proceedings.

And almost everything being exchanged in discovery is digital. There's no paper. It's zip drives and portal links and a million documents.

The proposed change flips the burden in a targeted way. If you've timestamped your contracts and records into Bitcoin as a standard business practice, and the hash of that document matches what's in the chain, the court treats it as presumptively authentic. The other side can still challenge it, that's why it's a rebuttable presumption, not an absolute one. But you've potentially cut out 80-90% of the discovery fight before it starts.

The three cases Bryan and Jacoutot cite in the comment as precedents, and I want to be precise here, these are cases named in the filed comment itself, which is the source to check for full citations, include instances of AI-generated video submitted as court evidence, metadata manipulation used to backdate digital records, and a case where a litigant used AI to alter an 8-second video clip that was subsequently discovered and resulted in a dismissal with prejudice.

That last one is the one that sticks with me. The litigant may have had a real underlying claim. It didn't matter. Dismissed with prejudice means the door is closed.

Bryan notes that the committee has been responsive. They told him and Rafael how they'd prefer to see the comment structured. That's not nothing. On the state level, every state has its own rules of evidence, and Bryan's view is that getting even one state to adopt this standard creates a working laboratory for the federal system to observe.

Screven County is already playing that role for Georgia at the county level. The goal is to get it statewide before 2026 voting concludes.

The DocuSign Problem and the Business Case for Doing This Now

The idea that got me most animated in this conversation was the DocuSign case, and it's simple enough to say in one sentence. You go to DocuSign or SignNow or any e-signature platform and say: I'll pay $5 extra a month to automatically broadcast a transaction that hashes every contract I sign into Bitcoin's blockchain.

That's it. That's the business model. The automation is already there. The protocol is already there. The cost is negligible.

The protection it provides in a dispute is enormous.

Bryan put it plainly: this is about building it into the normal course of business, not deploying it defensively after someone files a claim. If you're hashing because someone already challenged your evidence, you've lost the thread. The value of the timestamp is that it predates the dispute. A contract with a Bitcoin timestamp that was hashed at signing, before any disagreement existed, is a document where the authenticity question is already answered before anyone asks.

Even outside litigation, he makes the point that two parties with a reasonable disagreement about which version of a contract is current can just compare hashes. Skip the lawyers entirely. Go to the chain, look at what was hashed, resolve the dispute in a conversation instead of a courtroom.

Bitcoin companies should be leading on this. As Bryan put it: sell what we buy. If you believe in this technology enough to save your life savings in it, building your business records on its timechain is a low-hanging fruit that's already available.

The Regulatory Picture: What's Won and What Isn't

Bryan is genuinely optimistic about where things stand relative to five years ago, but he's not naive about what's still broken.

The win he's most pleased about: FinCEN's proposed rulemaking that would have classified basic good Bitcoin hygiene, fresh wallet addresses for each transaction, not reusing addresses, as a form of convertible virtual currency mixing. That rule would have put normal Bitcoiners on the same regulatory footing as mixers.

Coin Center filed a comment, others followed, and by Bryan's account, FinCEN retracted that rule. It's a real outcome from a grinding, unsexy process. Bryan's read: submitting comments works. It's not glamorous, but it changes what the government does.

The Blockchain Regulatory Certainty Act is the thing he wants next. It protects non-custodial developers from money-service-provider licensing requirements, essentially saying that writing code isn't the same as handling money, which is a reasonable position. The BRCA keeps getting bundled into larger market structure bills, most recently the CLARITY Act, which failed.

His argument is that if you stripped it out and ran it as a standalone bill, it has broad enough consensus to pass. Small win. Clear language. Gets Congress thinking about Bitcoin development as something other than a regulatory gray zone.

On the Bank Secrecy Act: I asked about outright repeal and Bryan laughed. Not happening. His floor is rolling back the Patriot Act additions and getting back to the pre-9/11 BSA. Even that's ambitious.

My position: at minimum, take the reporting thresholds that trigger suspicious activity reports and adjust them for inflation. Those figures trace back to the early 1970s. Pulling in everyone on transactions that were meaningful dollar amounts then and are trivial now is not law enforcement, it's noise. Bryan put it well: "The fact that we're using figures from 1971, 1973 is insane."

He's also clear that he personally believes mixers should not be prosecuted, a view he holds independently of Clark Hill, and I think he's right.

The Library of Congress and the Question Nobody's Asking

I want to close on the thing that's been bouncing around in my head since the conversation ended.

The Library of Congress holds hundreds of millions of documents. If you hashed all of them into Bitcoin's blockchain right now, you would have a cryptographic record of what those documents actually said at this moment in time. That's a firewall against future alteration, whether by a bad actor or by generative AI scraping the web and feeding modified text back into the historical record.

Bryan's comparison was the Library of Alexandria. That fire was an accident. What we're describing, AI companies training on historical content and then deleting or burning their training data, or internet archives being silently modified, is happening on purpose or through negligence.

You can't prevent a determined actor from altering a PDF on a government website. You can hash the real version before they do.

The window is open right now. Hash the documents as they actually exist, in their current state, before the AI models start rewriting the record. Give future historians something real to compare against.

Bryan said it directly: if we start hashing now, we're in decent shape. If we start in 10 or 12 years, we might be hashing a lot of fake stuff already.

I genuinely believe future historians will look back at the Bitcoiners who pushed this idea, who thought to use a "crazy" distributed timestamp protocol to anchor the digital record of civilization, and say they saved something important. That assessment follows directly from the trend line of what generative AI is doing to the concept of digital truth.

About Bryan Jacoutot

Bryan Jacoutot is a commercial litigator at Clark Hill, a full-service law firm with a Bitcoin and digital assets industry practice. He speaks on these topics in his personal capacity.

He has been working with Carlos Cordon and the Simple Proof team to apply the Open Timestamps Protocol to American election systems, business operations, and federal court evidence standards. Bryan and Rafael Cordon co-authored an 11-page comment to the Federal Rules of Evidence Advisory Committee proposing a rebuttable presumption of authenticity for Bitcoin-timestamped digital evidence.

He is reachable on X.

Sources mentioned

• Open Timestamps Protocol: the base protocol, introduced by Peter Todd, that Simple Proof builds on for Bitcoin-anchored document hashing
• Simple Proof: Carlos Cordon's implementation of the Open Timestamps Protocol applied to elections and business document verification
• Federal Rules of Evidence Advisory Committee (uscourts.gov): the body that received Bryan and Rafael Cordon's 11-page comment proposing a rebuttable presumption standard for Bitcoin-timestamped evidence
• Blockchain Regulatory Certainty Act (Congress.gov): legislation to protect non-custodial blockchain developers from money-service-provider licensing requirements
• Coin Center: nonprofit that filed comments opposing FinCEN's proposed rulemaking on address reuse and convertible virtual currency mixing
• AI Safety's Blind Spot: Closed Labs Can't Audit Themselves (TFTC): on the structural problems with centralized AI verification and why independent anchors matter

Watch the conversation

Timestamps

• 0:07 - Intro
• 0:34 - Bryan's Background and Bitcoin Journey
• 14:23 - AI Video and Why Timestamping Is Wildly Underappreciated
• 26:20 - Sponsor Reads
• 33:35 - AI Clogging the Courts
• 42:14 - DocuSign and Scaling Timestamping
• 49:00 - Low-Hanging Legal Wins: BRCA and FinCEN
• 52:26 - Society-Wide Implications and Truth Decay
• 58:07 - Calls to Action and Where to Find Bryan
• 1:00:04 - Closing

Sponsors

• Cash App: For a limited time, new customers can get $21 added to their balance. Just use code TFTC10 when you sign up, and send at least $5 to a friend in the first two weeks. cash.app/app/TFTC
• Square: For up to $200 off eligible Square hardware. square.com/go/tftc
• Bitkey: Use code TFTC10 for 10% off the new Bitkey. bitkey.world
• Aven: aven.com/bitcoin
• CrowdHealth: joincrowdhealth.com/tftc
• Simple Mining: simplemining.io/tftc
• Salt of the Earth: drinksote.com/tftc