Can we Prove we are Human in a World of Machines?

If Web2 achieved a wholesale datafication of life, it also removed control over that data from the human bodies who generated it. The recent rise of agentic AI has renewed focus on the value of smaller AI models that might be controlled by their users rather than centralized tech monopolies. Thanks to a combination of greater data literacy, decentralized AI, and proof-of-personhood (PoP) the possibility now exists to restore human agency at a time when national governments are threatening their own citizens’ digital sovereignty.

For her work, Heart Space (2024), Krista Kim partnered with Tenbeo to transform viewers’ heartbeats into a real-time generative experience that provokes discussion about proof-of-humanity protocols. Here, the artist hosts Tenbeo’s CEO Athalis Kratouni in conversation with data scholar Michael Clark on the prospects for a universal authentication technology.

Krista Kim: Michael and Athalis, what can you tell us about your work in the field of data sovereignty?

Michael Clark: Much of my work today centers on turning data into a real-world asset. In 2024, I worked with Dubai’s financial regulator to draft one of the first papers to explore how data can be treated as an asset in regulation. Now I’m helping governments and organizations globally embed this concept into policy and infrastructure.

We’re at an inflection point where AI is holding up a mirror to our society, exposing the gaps in how we learn, govern, and grow. But it’s also an opportunity because we now have the tools to rebuild our systems around data, identity, and intelligence, and unlock their real value in the process.‍

Athalis Kratouni: I am the co-founder and CEO at Tenbeo. We’ve developed a biometric technology that allows people to authenticate their identities using their heartbeat as a private key. The way our heart expands and contracts generates electricity that is translated to an electrocardiogram (ECG). We leverage machine learning to train a model from it, which becomes a person’s unique heart signature.

What we’ve managed to do — and we have a patent for this — is to enable each person to generate a cryptographic key from their unique heart signature, enabling the most intuitive user experience for people to securely access and encrypt their data.

Our first use case was biometric access control, but now we’re also developing products that we’re positioning as anti-phishing tools to verify the authenticity and humanity of the sender of digital content.

‍KK: Tenbeo is my core technology partner for Heart Space (2024), a biometric AI immersive artwork that premiered at Art Dubai 2024 and Noor Riyadh festival, with an upcoming show planned for London this July. Heart Space captures each participant’s unique heartbeat signature in real time and transforms it into a generative visual experience — a living portrait of human emotion and presence.

I chose Tenbeo’s technology because it provides a secure method of digital identity verification. In Heart Space, the heart signatures of four individuals are translated into unique waveforms displayed on an LED wall as an immersive experience. Each waveform is linked directly to the participant’s heart signature, and the color reflects their real-time emotional state. As of this year, visitors will be able to purchase their personalized Heart Space Signature, which can then be used beyond the art experience and serve as a digital watermark on emails, and digital platforms — offering a meaningful, sovereign proof of personhood. 

Tenbeo’s heart signature protocol is essential because, with the rise of AGI (Artificial General Intelligence), there is an existential risk that our digital identities and intellectual property could be overtaken or misused. We believe that combining proof-of-humanity protocols with widespread data literacy is crucial to establishing a secure foundation that ensures that human agency, creativity, and dignity are protected and can thrive in the age of AI.

‍MC: We look at data today as transactions rather than people’s stories and memories. I’m working with standards bodies to value data [which] is a representation of our culture as well as a unit of potential. I think of projects such as GeoCities and MySpace that were both deleted. If you add all that up, it’s like deleting the New York Public Library. During the COVID-19 pandemic, The National Archives in the UK struggled to capture what was happening in real time because many departments were routinely deleting data or not preserving it properly. Data touches every aspect of our lives but our regulations and institutions haven’t caught up. We’re still thinking about data privacy rather than focusing on ethics, trust, and security.

It’s 2025, and none of us has a unified view of our own health. I should be able to connect my health data with environmental data and know not to enter a building if allergens are present. 

I should be able to share my diabetic data with someone else and learn from how they manage their condition; maybe even discover a cure in that dataset. Eventually, our personal collections of data will power our own intelligence systems. But right now, we’re giving away data freely without getting back any utility. The world is regulating how AI processes data but not where that data originated, which is why identity also plays such a key role.‍

AK: The story behind Tenbeo comes from our own experiences with identity. I was born in Burundi, a small country in East Africa, and I found myself with three different identity cards. My Burundian ID is a piece of blue paper with a picture attached to it. If I were to put someone else’s picture on it, they could easily claim to be me in Burundi. Then I have the Greek identity card, which is a piece of plastified blue paper, and a Belgian residency card that has a chip.

What we can access in the digital world is defined by the identity card that we hold. This is why we wanted to develop a universal authentication technology based on the most intrinsic feature of human life: the heartbeat. Whether a baby is born in South America or the US, it all starts with a heartbeat. We wanted to bind verifiable credentials to one’s unique heart signature, allowing people to maintain control over their credentials — records and grades — that they produce over a lifetime. Being able to bind these to my own heart signature would enable me throughout my life, especially when moving between countries, to share data intentionally by providing heart-based cryptographic approval. 

‍KK: If you look at the human body, it is made up of atoms; if you look at the digital body, it is made of data. But in the age of AI, identity is no longer confined to the physical realm — it spans the material and the virtual. The ancient Greek axiom “know thyself” must be reinterpreted for this new era, where the self exists across dimensions. To truly know ourselves now means understanding and protecting our data. Agency, authorship, and ownership of that data are fundamental rights and they must be recognized in cultural frameworks and enshrined in government policy.‍

MC: We’ve been working with multiple governments and institutions globally to deploy the concept of data as an asset. One government initiative is focused on using data to help revive the economy in three key sectors: medical, energy, and agriculture — enabling data sharing between farmers and the creation of data marketplaces to establish a new asset class from the ground up. 

Another government has expressed interest in protecting health data by making it fully available, recognizing that once it is connected and protected, it can serve as a foundation to transform other industries. A separate project is concerned with designing an on-chain city powered by connected, trusted data. Across all of these discussions, one thing is clear: the real value lies in the data because, without data, there is no AI. Sovereignty means having the right to decide how your data is used. 

Recognizing data as a national asset means acknowledging that it is as vital to daily life as electricity, and should be managed with the same care. Yet today, some nations are giving away their data to AI companies — essentially handing over the keys to their digital heritage, cultural identity, and future economic engine. 

‍KK: Today’s AI ecosystem is dominated by centralized power structures: a handful of corporations like OpenAI, Google, and Anthropic that control the most advanced models, dictate access, and extract massive volumes of data from users. It resembles the Netscape-AOL era of the internet — an early, walled-garden phase where power is concentrated in the hands of a few gatekeepers.

But a new paradigm is forming: decentralized AI. Open-source models like DeepSeek are early signs of this shift. While still trained using centralized computing infrastructure, DeepSeek models can be freely downloaded, modified, and run locally — challenging the monopoly of closed systems. MIT’s Decentralized AI initiative is one of the leading academic efforts exploring how distributed architectures, federated learning, and cryptographic privacy can provide alternatives to centralized AI infrastructure. These efforts signal a growing recognition that the current trajectory of AI — centralized, opaque, and extractive — is neither sustainable nor equitable.

Centralized AI demands blind trust in black-box systems governed by corporate interests. Decentralized AI, by contrast, offers transparency, self-ownership, and digital agency. ‍

MC: In the past, instead of curating what models need, we’ve trained them on the entire internet, much of which is duplicated, unreliable, or retracted content. This reflects an outdated mindset whereby more data was considered better simply because we lacked proper valuation methods. A smaller model trained on high-quality, accurate data has proved to be more effective. The recent drop in GPU-related stocks is part of the realization that we may not need massive models after all. Agentic AI was a natural next step, emphasizing smaller, purpose-built models. The next shift will be toward personal AI that lives on your device. The question is: what data is powering that AI? Is it owned and controlled by you, or is it managed by someone else?‍

KK: A personal AI model running locally on your device becomes more than just software, it becomes your digital twin: your coach, confidant, and optimizer. It understands your biomarkers, life history, family background, and your unique learning profile. By aggregating this deeply personal data, your digital twin can represent you in the arena of AI: communicating on your behalf, advocating for your interests, and protecting you from exploitation by AGI, corporations, or institutional systems. It is your intelligent shield and trusted ally in the digital age.‍

MC. The concept of a “super agent” has long been central to my thinking about personalized AI and data sovereignty. What sets my vision apart is where that agent lives. Most of today’s models imagine it running in the cloud, not on data you truly own. But there’s an opportunity here: imagine a snapshot of your intelligence — secure, private, and fully under your control — running directly on your device. Thanks to technologies like secure enclaves, that future is already possible. This moment is a catalyst not only for how we build AI, but for how we embed trust, fairness, and real value into its foundation. And ultimately, how we ensure that those who contribute the most data are the ones who benefit the most.

‍KK: What is the difference between Sam Altman’s Worldcoin project and Tenbeo’s approach to proof of personhood?‍

AK: Worldcoin has been banned by certain countries, including Kenya, because the local government understood that Worldcoin was collecting the irises of hundreds of thousands of people with a promise of giving them tokens that have no real utility in the local economy. When you do an iris scan, you’re also capturing data about the retina, which can be used for scientific research to gather information about people’s health. The Kenyan government sensed it was a big data collection operation based on a false promise of giving money to the local population.

‍Technology is like a knife — it can be used to cut bread the same way it can be used to kill. With biometrics technology as with other technologies, it all comes back to the ethics of the people running the company. 

In comparison to how Orbs (biometric verification devices) are being operated in specific locations, the difference with Tenbeo is that every single person can own their own ECG sensor and access it via their smartwatch or a dedicated Tenbeo device. In this way, a person can use their heart-derived private key at any time to securely authenticate, sign an email, or provide proof of personhood while uploading content online.

‍MC: When we talk about our right to data, the conversation often shifts straight to money: what’s my share? How do I get paid? And while those are fair questions, the danger is that we end up replicating the same extractive thinking where power remains centralized and short-term gain trumps long-term transformation. 

‍What this is really about is human potential: what can we do and what can’t we do today because we don’t have access to it? And what does it mean when we can’t prove who we are? 

We’ve ended up here because technology has evolved in two waves. The first wave gave us centralized platforms — social media, big tech ecosystems — that made life more connected but also concentrated control. The second wave brought decentralized technologies like blockchain, offering more control to individuals but often without clear direction or accountability. Instead of choosing one over the other, we can take the best of both — trusted infrastructure and personal empowerment — and make a conscious decision to go in a better direction. For the first time, we have the chance to decide how we want this technology to work for us and not just around us.‍

KK: As autonomous AI agents become more powerful, the potential for exploitation — intentional or emergent — grows exponentially. Without protections like Tenbeo’s heart signature protocol, decentralized AI infrastructure, and enforceable data sovereignty, these systems can operate with criminal efficiency: replicating, impersonating, and monetizing creative works without consent. These risks extend beyond deepfakes and identity theft to include algorithmic erasure and systemic manipulation. We must urgently establish ethical frameworks, decentralized controls, and government policies that defend creators from exploitative, agentic AI systems acting without accountability.