How our Seevik Net aims to radically mitigate social media disinformation and generative-AI-powered chatbots.

Last June 6th, we published on Linkedin, and on our blog, our first post in a long time about AI, AGI and ASI, detailing our position on their risks and opportunities, and our foreseen role to substantially helping foster positive scenarios, via our Trustless Computing Certification Body and Seevik Net Initiative, from mitigating its impact of social media disinformation to that of AI existential risks.

A long-time observer of our initiative, Thomas Dübendorfer, replied to such Linkedin post asking us to write a follow up post to answer how our Initiative plans to "ensure that a specific human is on the other side instead of an AI-powered chatbot".

Here it is. (Reading first the 700 words Linkedin post mentioned above is highly preferable.)

Currently, dominant social media networks do not require strong initial identification of users when registering, nor require strong authentication for each user when logging in. That is due to the fact that netizens have gotten used to the “feature” of pseudonymity and anonymity on such networks, which is a democratic right in western democracies (albeit largely an illusion today for nearly all users).

Hence, nation states and shady entities, state connected and not, run armies of automated and semi-automated chatbots to manipulate large swathes of the population and especially of influential users, such as swing voters and influential persons.

With the recent advances and widespread availability of increasingly powerful “Generative AI” systems, like OpenAI’s ChatGPT, it is rightly feared that those entities will be able to "turbo charge" their nefarious activities.

How can we enable law-abiding citizens much stronger initial identification, stronger authentication while at once retaining and improving the ability for users to reliably participate in social media in both anonymous and pseudonymous ways compatible while enabling law enforcement to prevent their at-scale or grave criminal abuse?

There is not an app for that, nor a new law or regulation that will fix it.

The solution goes through a (1) hardware solution, through (2) the creation of a new inter-governmental certification body to extreme levels of trustworthiness of both a complete end-to-end IT service and of in-person procedural mechanism for legitimate lawful access, national an international.

A policy push, at the national and inter-governmental levels - to progressively mandate those for client devices and dominant social media, initially for millions of the most sensitive users and then for all - is preferred but not requires as (1) and (2) can be fully enacted within current national and international laws.

Let’s see why and how.

Tens of millions of the most sensitive users are currently required by their institutions to carry external multi-factor authentication keys, like Google Titan Keys, to achieve substantially higher levels of authentication.

These are cumbersome to use and set up, but quite effective. Yet, they do not protect against a malware on the device that can take control of it - as it happens to millions globally - other hacks and especially man-in-the-middle attacks whereby the user thinks he is talking to a given person, reading a certain content on social media, or chatting with ChatGPT, while instead he is talking to a human impostor, automated AI-powered chatbot, or a combination of them.

At the Trustless Computing Association, we are building a new inter-governmental IT certification body for client and server IT systems for human-to-human and human-to-and-through-AI communications - the Trustless Computing Certification Body (or TCCB) - that achieves levels of confidentiality and integrity radically higher that state-of the-art - while ensuring legitimate lawful access, national and international, via in-person procedural mechanisms,to reconcile digital civil rights with the needs of global public safety in an increasingly dangerous world.

Meanwhile, its spin-in startup TRUSTLESS.AI, is building a complete TCCB-compliant IT system, infrastructure and platform - Seevik Net -that includes  a private cloud and 2mm-thin, standalone, ultra-secure and minimalistic mobile phones - embedded in the back of the user’s smartphone or carried in custom leather wallets - that act both as ultra-secure authentication device as well as standalone connected client computing devices.

Such TCCB-certified Seevik devices will radically mitigate the risk of man-in-the-middle attacks by having a TCCB-certified VPN chip on the client device and on the server-side, and on TCCB-certified cloud services (such as ChatGPT or other server side service).

Seevik Net, with its much stronger endpoint security and multi-sensorial behavioral biometrics will provide radically higher resistance to the malware attacks mentioned above, even against state-grade spyware, and significantly higher authentication and non-repudiability (for digital transactions), as well as more reliable digital watermarking for any content shared by the user.

Seevik Net will enable the user to engage in discussion using one or more pseudonyms, in radically more secure ways than today, but which will be reversible in case he/she becomes suspect of a crime through a civilian due process.

The cost of Seevik devices at scale will be of about $200, which is significantly higher than a typical external multi-factor authentication key that runs at about $50, but the additional features the Seevik device will bring incomparable additional utility. An app store will be available with anyone being able to write TCCB-compliant apps for those devices.

While using only components that are open source, or at least publicly inspectable in their source designs, and implementing state-of-the-art end-to-end and quantum-resistant encryption, all user data will be saved on a TCCB-compliant ultra-secure, distributed, multinational private cloud - accessible only through the in-person approval of 5 random-sampled citizens and/or an international judicial board.

This solution will initially be available to the most influential and targeted persons in our democratic societies, like elected officials, journalists, activists, and top businessmen, and mandated for some, like elected officials.

Shortly after, Seevik devices will be given out to all citizens of a given nation, with a combination of subsidies for the poorest, and cross-subsidization of the cost of the device with revenue-share generated by non-included apps and services.

It will be a sort of ultra-secure, mobile and multi-national version of the Minitel, the hugely successful PC/terminal that the French government distributed to its citizens in the 80s in millions of units, built by several private companies in different models, and with tens of thousands of private apps.

Join us in trying to turn AI into the greatest invention that humanity has ever made, while staving off AI dystopias and catastrophic risks!