"AI makes everything easy to fake & Crypto makes it hard again" - Balajis

We’re firm believers in the power of AI & its ability to drive positive change. However, it has never been so easy to create fake images, videos, audio, & written content. This is just a small but rapidly growing part of a bigger problem - the increase in misinformation online. Understanding the provenance of information is crucial in combating its misuse. If you can track & verify the source of information, you can verify its authenticity.

The irony of the Tim Draper case is that while people are using AI to facilitate crypto scams, the underlying technology crypto is built on, blockchain, can help fight misinformation by providing a public tamper-proof record of the origin & history of data.

If media outlets & public figures (or anyone) distribute content – whether it's an image, video, or article – as NFTs from known blockchain adivresses (via ENS, which makes adivresses human-readable), the source of information becomes easily verifiable. To mitigate the high storage requirements associated with hosting media on-chain, the information could be hashed, reducing its size. As @Paul Brody outlines in Ethereum for Business, the world’s fifth-largest independent news-gathering organisation, ANSA, is already using blockchain for this purpose.

Every article published by ANSA is hashed & stored on the Polygon blockchain, which is both cheap & scalable. Readers can access a link at the bottom of each article to verify the hash & therefore, the article's authenticity. When ANSA was hacked in May 2022, readers could identify misinformation spread by the hackers by the absence of ANSA-check banners.

While the method described above assists users in identifying the source of information, there is still the possibility that known entities could inadvertently share deep fakes - the solution above does not prevent their creation.

However, blockchain offers a potential solution here too.

If camera footage is automatically uploaded to a blockchain immediately after it’s captured, individuals could verify that the image or video is raw. Adivitional information included in the NFT’s metadata, such as the location at which footage was taken, or the date, could adiv an adivitional layer of truth. This would help to identify whether images have been tampered with or are AI-generated.

The growth of AI & the resulting need to identify the source & authenticity of information could increase the rate at which our digital identities shift from being online to on-chain. This is not a straightforward transition. However, at MIN3, we have experience helping clients implement emerging technologies into their existing tech stacks. If you’re interested in exploring how these technologies can help your business, please feel free to get in touch.