Using VIDT’s timestamping tech to verify AI models

By now it is very clear that AI is the future in the technology industry. With the great success of OpenAI’s ChatGPT, and the planned releases of the AI tools of other tech giants, a lot of growth has happened in these markets. That growth is seen in all aspects: financially, the usage of the tooling, but unfortunately also in the amount of fake news and wrong information that is being generated and used by the AI models.

The challenge of ensuring the verifiable authenticity of data has been a central topic among members of the VIDT DAO ecosystem since the technology's inception. This proposal aims to address the hot issue of misinformation and fake news in AI applications by introducing a new integration of the VIDT DAO timestamping technology. AI-Validator seeks to enhance data integrity and authenticity across AI platforms, offering a robust solution to one of the most pressing challenges in the AI landscape.

The Problem

AI models are essentially based on all the scrapeable information on the internet. This is fed into the learning models and then released to the public. All the big tech companies report that data and information is verified before it is fed into the models. But unfortunately, many times it has been proven otherwise.

Unverified facts are reported as verified, wrong statements are given as an answer to a question, or even new information is made up based on other answers. We understand that this is the case and that each individual instance cannot be prevented due to the sheer size of the AI tools. But, something has to be done before it is too late. Our future generations are growing up with this new technology, and we need to make sure that we are basing everything on facts and the truth.

The Solution - AI-Validator

We propose to develop a new tool; AI-Validator. With AI-Validator, facts and statements can be timestamped as truth by individuals and organizations which then can be fed to AI models to guarantee that the models are based on truth and science.

It would enable anybody to verify if a fact is true or false, and who has stated that is true or not. This could even be a source to be referenced to by the AI model in the chat itself. That way we are directly integrating the ‘Do Your Own Research’ (DYOR) mindset into the technology, which we believe is a very good foundation to build the tooling on.

AI-Validator could even be used to timestamp scientific papers and the results of research so that the verified information can be safely used, instead of using speculative conclusions based on the information.

FileValidator has proven that the VIDT DAO Timestamp technology is very scalable and anybody can develop use-cases on top of it. A-IValidator is the first of the next generation of integrators.

The Proposal

In order to develop this new technology, we are asking for a grant of 1.85 million tokens. This will go towards the development of the first version; a proof of concept to prove the validity of the technology.

And therefore, the grant goes towards:

• Design: designing the UI and UX for the dashboard and the pilot version of the AI-Validator tool. This will be used by researchers for example to validate the facts and information.
• Development: When the design is finalized, and our test users are satisfied with the results, we will start the development of the tool. This will start as a pilot (or a proof of concept if you will) to showcase the technology and the usage.
• Marketing: And of course, to grow the VIDT DAO ecosystem, we will be doing marketing of this new product (and integrator). This is also done to start reaching potential users of AI-Validator.

It is also important to note that the grant is used for the maintenance, and upkeep, of the tooling for at least a year.