Project summary

Before Day Zero (BDZ) is essentially an online 'living museum' designed to keep alive the experience of using the internet before AI gradually seeped into all areas. Our aim is to establish an open platform containing genuine and verified pre-AI media which we can mathematically demonstrate was created before generative models became widespread. Together with the files, we are also developing interactive areas known as 'Cognitive Sandboxes'. The concept is to simulate typical digital life including all its awkwardness, uncorrected errors, and the amount of thinking that was necessary back when auto-suggestions and LLMs began to smooth things over. It serves as a cultural snapshot for researchers, future students, and anyone who wishes to remember how human brains functioned when online without an algorithm taking on half the work.

What are this project's goals? How will you achieve them?

The goals are:

Create a genuine and immutable archive of material from before 2022.

We need a dependable method of storing authentic raw files, for example old low-resolution webcam videos, message board threads, and digitised VHS tapes, so that people in the future have a real ground-truth reference point showing what ungenerated media actually looked like. We intend to incorporate this feature into a simple and lightweight web application which operates via standard HTTPS and continues to function even if users are offline.

Create the sensation of 'digital friction'.

We would like people to have the real experience of using older technology, which involves using browser tools without autocomplete, carrying out manual searches with no suggestions, and browsing static pages without being prompted by feeds. In order to obtain authentic material, we will get in touch with digital libraries, community web archivists, and humanities labs. On the technological side, we will verify the file dates by means of public decentralised timestamping (such as RFC 3161) and compare them with the existing pre-2022 snapshots from the major web archives.

Provide researchers with a pure starting point for unassisted human thinking.

People who study cognition and those who work on AI policy should understand the ways in which humans naturally deliberated, wrote, made mistakes, and edited before generative AI was introduced into the process. We will convert our simulation setups into open-source modules and SDKs so that teachers and university laboratories can easily incorporate these sandbox tasks into their classroom research.

How will this funding be used?

For 45 per cent of the time the focus was on core web development and system architecture, including the establishment of the main web platform, the backend, and the sandbox environments.

For 25 per cent of the budget: reliable and secure long-term storage to ensure that the preserved files are safe, easily accessible and quick to retrieve.

The 20% Curation and Legal Verification Pipeline involves carrying out the timestamp checks, examining the collections, and managing the content permissions.

10 per cent on user testing, outreach and documentation: carrying out trial runs with students and researchers and preparing clear guides for open-source users.

Who is on your team? What's your track record on similar projects?

I'll start it off myself and then employ team members to carry it out. As an Assistant Professor of Computer Engineering at the Middle East Technical University Northern Cyprus Campus (METU-NCC), I obtained my PhD in Computer Science and Engineering from Kyungpook National University in South Korea. My professional experience has been in the areas of networked communication systems, edge architectures, IoT, and the application of machine learning; I also have more than 35 publications in IEEE transactions and high-impact journals, have co-authored some patents, and have written a textbook on network programming.

What are the most likely causes and outcomes if this project fails?

The project has the possibility of failing as a result of low ongoing retention, since users may go to the platform out of a sense of nostalgia and not incorporate it into their regular educational or personal routines. In order to counter this, we should concentrate more on the pedagogical tools available for schools, the inclusion of research citations, and the API and embed tools rather than simply depending on traffic generated by novelty.

As AI is also advancing and the number of submissions can surge, this may result in contamination of the archives and bottlenecks when verifying them. In order to deal with this, we should introduce strict algorithmic cutoffs based on the existing public web archives (the Internet Archive's Wayback Machine snapshots from before 2022, along with EXIF/hash verification).

How much money have you raised in the last 12 months, and from where?

We have very recently secured a number of international research funding awards through competitive processes, among them grants from the International Science Partnerships Fund (amounting to £80,000), the National Research Foundation of Korea (£450,000), the Québec Merit Scholarship programme (£35,000), and the AdımODTÜ research programmes.