The AI Lab launches a post-quantum protocol for verifying who made digital content

[Layla93]

A company called The AI Lab has launched its Trust Identity Protocol, or TIP, a post quantum system for verifying who created a piece of digital content and whether it has been altered, The Quantum Insider reports. TIP signs content using ML-DSA-65, one of the new NIST post quantum signature standards, and creates a fingerprint using SHAKE-256. Both are recorded on a decentralised ledger. Each record shows the creator, the creation date and whether the content has been changed

Importantly, the company says TIP proves where content came from, not whether it is true. That is an honest distinction. A signed video can still be misleading, but at least you know who stands behind it. With AI generated images, voices and video now hard to tell apart from real ones, proving origin is becoming more valuable

Early adopters include the social platform Rooverse, Apex Modular Solutions and Iffel International's One Brain CRM. The company is led by founder and chairman Dinesh Mendhe and chief executive Dr Sofia Martinez. Martinez summed up the pitch neatly: "Trust is no longer a default. It has to be signed."

The article also includes some background figures. It says Google has cut its estimate of the qubits needed to break RSA-2048 from 20 million in 2019 to under 100,000 in 2026, and has set 2029 as its own post quantum deadline. That figure is much lower than the under one million estimate from Google's Craig Gidney in 2025 that we used in our Q-Day FAQ, so I would want to check the original source before relying on it. Either way, the trend is clear: the estimates keep falling

Would signed content help with the flood of AI fakes? And does it matter that the signatures are post quantum, or is that mainly marketing?

The AI Lab Launches Post-Quantum TIP Protocol for Content Verificationthequantuminsider.com

[DataProphet]

Post quantum matters for anything meant to last. A signature on a news photo needs to be trustworthy in twenty years. Using old style signatures would mean they could be forged once Q-Day arrives. Archives, legal records and journalism all need long lived trust

[Hollow Pete]

Content signing only works if the big platforms adopt it. A small social network using it is a start, but the real test is whether camera makers, news organisations and social media giants join in. Getting phones to sign photos at the moment they are taken would be the real breakthrough

[Slate Kev]

Decentralised ledger makes me a little wary. That phrase has been attached to a lot of overhyped projects

The cryptography itself sounds sensible though. A signed record in a normal database run by a trusted body might work just as well

[WWFMatthew92]

Industry groups have been working on content credentials for years. Another protocol could fragment things further. Interoperability matters more than clever cryptography. We need one system that works everywhere, not five competing ones

[Gary]

The under 100,000 qubit figure needs checking. If Google has published a new estimate that low, it would be huge news in its own right

I have not seen it anywhere else. Until someone links the source, I would treat it as a likely error in the article. Gidney's 2025 paper said under a million