The AI Lab

The AI Lab

We build open infrastructure to make truth verifiable. Your name belongs on your work. Biometrically sign anything you publish on the internet.

Claim your content before anything else does. TIP Protocol proves who made content and how.

10/08/2026

A quiet milestone for a loud mission.

AI Trust Councilยฎ is now a registered trademark.

On October 6, 2026 the USPTO registered it to The AI Lab on the Principal Register, U.S. Reg. No. 8483190.

When we built the Trust Identity Protocolยฎ, one principle was clear from day one. Code and cryptographic proof are half the work. The other half is human judgment, policy, and accountability across industries.

So we formed the AI Trust Council, an independent roundtable of policy experts, enterprise leaders and ethics advisors who take on the hardest questions in AI provenance and governance.

The trademark is a checkpoint for our team. Trust is infrastructure, and we are building it to last.

Thank you to our advisors, Council members and team, who give their expertise to keep people in charge of AI.

The work continues.

Meet the Council https://lnkd.in/eXUJW45Q

โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”
๐Ÿ”ต AA: AI-Assisted ยท Protected by proof on TIP
https://findtip.org/?c=AA-1c0042483aef02-ed1c
โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”

10/05/2026

In 2019, breaking RSA-2048, the encryption that still protects much of the internet, was estimated to need 20 million qubits. In February 2026, the Pinnacle Architecture estimate put it below 100,000.

That is 200 times smaller in seven years. A Google Quantum AI estimate had already dropped it under 1 million in May 2025. Google and Cloudflare have both set 2029 as their post-quantum deadline, and NIST proposes deprecating RSA-2048 after 2030 and disallowing it after 2035.

AI is part of why this is moving fast. Google Quantum AI and Google DeepMind used AI to improve quantum error correction with AlphaQubit in 2024. My view is that we now have roughly two years to prepare. The Global Risk Institute is more cautious and puts the chance of a code-breaking quantum computer within ten years at 28 to 49 percent. Both timelines lead to the same to-do list.

Four things worth doing this quarter, whatever you build.

Step 1. List every place you rely on RSA or elliptic-curve signatures, such as software updates, signed documents, logins and certificates.

Step 2. Ask each vendor for its post-quantum roadmap and a date, in writing.

Step 3. Start with what must stay trustworthy for years, like contracts, archives, firmware and anything published under your name.

Step 4. Test ML-DSA (NIST FIPS 204) in one real signing pipeline before a deadline forces it.

For content, we built the Trust Identity Protocolยฎ (TIP), and it is available now. TIP fingerprints each piece of content with SHAKE-256 and signs it with your key, which you unlock with a passkey. Each signed file becomes a Verifiable Asset with a public record of who made it, when, and whether it has changed since. It does not prove the content is true, and we say so plainly.

Three companies are already building on it. Ross Thorpe leads Rooverse, Issa Nesheiwat leads Apex Modular Solutions, and Hema Dey leads Iffel International Inc., whose One Brain is adopting TIP. Ross and Issa are Founding Members of the AI Trust Councilโ„ข and Hema is a Council Advisor, so we say that openly too.

Where does your organization still depend on RSA signatures? Tell me below and I will reply to every answer.

10/05/2026

๐ŸŒบ Niyam Bhushan helped lay the foundations of the Indian-language fonts and encoding that now run on Android, Apple platforms and the web. Today he joins the AI Trust Councilโ„ข as a Member.

Niyam has worked across design and technology for more than four decades, in over 20 industries, on engagements involving Google, Apple and Adobe as well as government institutions, startups and universities. From 2021 to 2023 he worked on health-AI initiatives for Google addressing population health for more than 80 million people. In 2020 he co-founded Pruve Systems, a self-sovereign identity venture with a digital signature designed to bridge literacy barriers.

He practises Privacy by Design, co-founded the India Linux User Group Delhi, helped launch Creative Commons in India, and through DesignRev.in, with support from nasscom, has mentored more than 300 Indian startups.

The AI Trust Councilโ„ข governs the Trust Identity Protocolยฎ (TIP), an open standard that records who made a piece of content, how and when, and whether it has changed since. ๐ŸŒบ Niyam brings his identity and privacy work to that job.

๐ŸŒบ Niyam Bhushan, welcome. I am glad you said yes.

If Niyam's work has touched yours, from the Hindi on your phone to a startup he mentored, say hello below.

Last month, Hollywood actors appeared in CNN videos attacking the Democratic Party ahead of the US midterms. The actors ...
10/03/2026

Last month, Hollywood actors appeared in CNN videos attacking the Democratic Party ahead of the US midterms. The actors never said those words. CNN never made those videos.

Researchers at Graphika and the ISD (Institute for Strategic Dialogue) traced them to Matryoshka, a Russian campaign also known as Operation Overload. It took real clips, including Cameo recordings and an old Julia Roberts ad for Gucci, cloned the voices, moved the lips and stamped a CNN logo on top.

Emma Caulfield, one of the actors, told AFP, "I feel violated."

A logo proves nothing. Anyone can copy one in seconds. If this can happen to famous actors and a global news network, it can happen to any creator, any business and any election.

That is why we built the Trust Identity Protocolยฎ(TIP). A verified person gets a TIP-ID. Everything they publish can carry a CTID, a signed public record of who made it, when, how it was made, and whether anyone changed it. Anyone can check it for free. We call that a Verifiable Asset.

A CTID will not tell you whether a claim is true. It tells you who stands behind it and whether it was altered, and every one of those fake CNN clips was missing exactly that.

On Tuesday, October 20, we launch TIP at OC Tech Week in Irvine. Creators from Orange County, Las Vegas and Silicon Valley, come get your TIP-ID and sign your first piece of content with our team beside you. It is free, from 12 to 4 PM PT at EvoNexus.

Would you have spotted the fakes? Watch the video and tell us in the comments. The registration link is in the first comment.



Welcome to the AI Trust Councilโ„ข, Mostafa ElkordyMostafa is Senior Cybersecurity Solutions Officer at UNICC, the United ...
10/02/2026

Welcome to the AI Trust Councilโ„ข, Mostafa Elkordy

Mostafa is Senior Cybersecurity Solutions Officer at UNICC, the United Nations International Computing Centre. There he leads pe*******on testing and secure communication and cryptography work, and he drives open source adoption across United Nations agencies as Co-Chair of Open Source United. He is also the founding author of OSARA, the Open Sovereign Agent Reference Architecture.

His path to UNICC ran through the Air National Guard and 18 years at United Nations Population Fund (UNFPA), more than 25 years in all spent securing the systems that institutions depend on. He holds an MBA and an MS in Software Development and Engineering, and he is CISSP and GCSA certified.

The Trust Identity Protocolยฎ is open, cryptographic infrastructure for proving who made a piece of content, and it has to earn the trust of many institutions at once. Open code, strong cryptography and shared infrastructure are the problems Mostafa has worked on throughout his career. We are glad to have his counsel.

Read his profile at https://lnkd.in/g6ZSDP9E

Visit The AI Lab at:
https://lnkd.in/eNc6iYkD

Follow us on:
LinkedIn https://lnkd.in/gPRm9K5T
X https://x.com/theailaborg
YouTube https://lnkd.in/gWPfq-8e
Instagram https://lnkd.in/g6uCYHrU

09/28/2026

Today we welcome Prof. (Dr.) Ajay Kumar to the AI Trust Councilโ„ข.

He is Vice President of the IETE, and Professor and Head of Computer Science and Engineering at RTCIT Ranchi.

The Council governs the Trust Identity Protocol, the open standard that answers one question about anything published online. Who made this, and when.

IETE has been training India's electronics and telecommunication engineers since 1953, and the engineers who will build on TIP in India are in its classrooms now.

Visit The AI Lab at:
https://lnkd.in/eNc6iYkD

Follow us on:
LinkedIn https://lnkd.in/gPRm9K5T
X https://x.com/theailaborg
YouTube https://lnkd.in/gWPfq-8e
Instagram https://lnkd.in/g6uCYHrU
Facebook https://lnkd.in/g5Zn_QKe

Ross Kathleen Hema Niamh Tory Gareth Nate Tricia Carolyn C. Alan Cyrus Joshua ๐ŸŒบ Niyam Garry Nikhil Bill Issa Mark Chris Vinh Sofia Vladimer Matt Ram

OpenAI built a tool to detect AI-written text. It caught only 26 percent of it. They withdrew it in July 2023.That is th...
09/25/2026

OpenAI built a tool to detect AI-written text. It caught only 26 percent of it.

They withdrew it in July 2023.

That is the whole point. If the companies building these models canโ€™t reliably identify their own AI-generated content, how can we expect anyone downstream to do it? And that gap is getting wider. Newsguard counted 3,749 AI content farm sites by June 2026. At some point, trying to detect everything after the fact will not be sufficient.

So the Trust Identity Protocolยฎ goes the other way. Instead of guessing after the fact, a verified human signs their work at the moment they publish it, and declares how it was made. Original human. AI-assisted. AI-generated. Mixed. The record is permanent, nobody can quietly edit it later, and checking it is free and open to anyone.

Rishabh Narayanan and Siddarth Narayanan of Tondova wrote this one for us. It is the clearest explanation of the point of TIP I have read, and it is not written for engineers.

Read more at: https://lnkd.in/gnVk-B88

https://findtip.org/?c=AA-c60e8abb3f129e-c4d4
โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”
๐Ÿ”ต AA: AI-Assisted ยท Protected by proof on TIP
https://findtip.org/?c=AA-c60e8abb3f129e-c4d4
โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”

09/22/2026

On 17 September, the European Commission published its proposal for the EU KIDS Act. Most of the attention has focused on the proposed age rules - children under 13 would not have social media accounts, while 13- and 14-year-olds could have limited accounts set up by a parent or guardian, with parental controls, contact restrictions and a daily time limit of no more than one hour. From 15, young people could create their own accounts on the services covered by the proposal. The rules also extend beyond social media, with provisions covering areas including online games, AI companions, general conversational chatbots, app stores and operating systems.

But there is another part of the proposal that deserves much closer attention.
Article 28(3) states that - โ€œAny age assurance measure shall be zero knowledge proof.โ€ That is not a recommendation. It is written directly into the operative text of the proposed Regulation.

That raises an important technical question - what would it take to implement age assurance in a way that proves an age threshold without identifying, tracking or profiling the person?

The Commission already has an age-verification architecture that incorporates zero-knowledge technology, so it would be wrong to say that such systems do not exist. The more important question is the gap between a technical specification and reliable, privacy-preserving deployment at the scale and scope contemplated by the proposal.

We read the proposal closely and examined what Article 28(3) actually requires, how the provision fits with the rest of the age-assurance framework, and the technical questions that still need to be addressed before this can work at scale.

The EU KIDS Act is a proposal, not law. The text will go through the EU legislative process, and provisions can change.

We have written up our analysis here:
https://lnkd.in/gCEYsqdf

Written with Ross Thorpe, Hema Dey and Sofia Martinez.

โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”
๐ŸŸข OH: Original Human ยท Protected by proof on TIP
https://findtip.org/?c=OH-3982a55ca3e37e-cfe1
โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”

In June, at United Nations Open Source Week in New York, I met Tricia Wang. We have been talking ever since.Today I am p...
09/18/2026

In June, at United Nations Open Source Week in New York, I met Tricia Wang. We have been talking ever since.

Today I am proud to share that Tricia has joined the AI Trust Councilโ„ข of The AI Lab as an Advisor.

Many people arrive at AI governance through theory. Tricia arrived through fieldwork. As an ethnographer at Nokia, she spent years with people in China and came back with evidence that low income buyers would stretch for a smartphone. She had 100 people. The company had millions of data points, and trusted the data points.

She gave that failure a name, quantification bias, the tendency to favor what is measurable over what matters. Her TED Conferences talk on it has been watched more than two million times.

Today she is Co-founder and CEO of the Advanced AI Society, which this week joined The Linux Foundation and opened its Proof-of-Control standard for public comment. One line of hers has stayed with me since we met:

"Verification cannot belong to the verified."

That is the whole argument for the work we are doing. Trust Identity Protocolยฎ proves who made a piece of content. Her work proves what an autonomous system actually did. Two layers of one idea: trust has to be checkable by anyone, or it is not trust.

Tricia, welcome. Thank you for standing with us early.

In June, at United Nations Open Source Week in New York, I met  Wang. We have been talking ever since.Today I am proud t...
09/18/2026

In June, at United Nations Open Source Week in New York, I met Wang. We have been talking ever since.

Today I am proud to share that Tricia has joined the AI Trust Councilโ„ข of The AI Lab as an Advisor.

Many people arrive at AI governance through theory. Tricia arrived through fieldwork. As an ethnographer at Nokia, she spent years with people in China and came back with evidence that low income buyers would stretch for a smartphone. She had 100 people. The company had millions of data points, and trusted the data points.

She gave that failure a name, quantification bias, the tendency to favor what is measurable over what matters. Her TED Conferences talk on it has been watched more than two million times.

Today she is Co-founder and CEO of the Advanced AI Society, which this week joined The Linux Foundation and opened its Proof-of-Control standard for public comment. One line of hers has stayed with me since we met:

"Verification cannot belong to the verified."

That is the whole argument for the work we are doing. Trust Identity Protocolยฎ proves who made a piece of content. Her work proves what an autonomous system actually did. Two layers of one idea: trust has to be checkable by anyone, or it is not trust.

Tricia, welcome. Thank you for standing with us early.

โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”
๐Ÿ”ต AA: AI-Assisted ยท Protected by proof on TIP
https://findtip.org/?c=AA-606ba3fbeb1a58-c4d4
โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”โ”

Address

131 Continental Drive
Newark, DE
19702

Alerts

Be the first to know and let us send you an email when The AI Lab posts news and promotions. Your email address will not be used for any other purpose, and you can unsubscribe at any time.

Shortcuts

Share