Blockchain technology is the answer to AI deepfakes

Last month, I watched “Mountainhead,” a film depicting a reality where deepfakes, phony videos created using artificial intelligence, cause mass hysteria. In the film, fake videos of bombs exploding and real-life leaders declaring war appear on social media, inciting real-life violence. As objective reality collapses, the world descends into AI-fueled anarchy led by an eerily familiar group of tech billionaires.

“Mountainhead” may be fictional, but deepfakes are very real. Over the last few years, deepfake attempts have increased along with their quality, placing an already fraught news environment at greater risk. Thankfully, there is a solution — blockchain technology — and in order to combat deepfakes, we will need to move on-chain. Fast. 

Blockchain technology refers to the digital “ledger” distributed between computers to record transactions. Without getting too technical, it allows for the secure and transparent sending of data independent of centralized institutions. While often associated with cryptocurrency, blockchain has many other uses, including insurance brokering, communications and education. Most importantly, data added to a “block” is unalterable and timestamped with a unique digital identifier. 

In practice, verifying authenticity through blockchain likely means viewing content on decentralized data storage ecosystems such as Arweave. Arweave, and applications like it, (BitTorrent, Sia, MaidSafe) would tag photos and videos with a unique identifier, which you would then add to a block. Anyone seeking to verify the authenticity of the content would cross-reference the identifier with the one added to the blockchain. Video sharing platforms will eventually do this automatically, comparing identifiers and adding visual proof of authenticity. 

This cannot come soon enough, as deepfakes are already challenging our perception of reality. Many are relatively harmless — Taylor Swift selling cookware, Eminem performing live or Jerry Seinfeld playing the lead role in “Pulp Fiction” to name a few — but some are beginning to cause harm. In May of 2023, videos circulated on X showing the Pentagon exploding, causing the stock market to briefly dip. Law enforcement immediately clarified that the videos were fake, but the scammers’ short-lived success shows that fake videos have real consequences. 

Take war, for example. In 2022, a video went viral deceptively showing Russian President Vladimir Putin ordering a full withdrawal from Ukraine. The video was completely fabricated and immediately followed up by similar content depicting Ukrainian President Volodymyr Zelensky surrendering to the Russian invaders. If both leaders had verified their official statements on-chain, this dangerous confusion would have been avoided. Instead, entire populations were left to speculate whether their country was still at war. 

This is why it is not enough to label deepfakes as “misinformation” and move on. AI-generated videos are insidious because they are true in a way that incorrect statistics or false claims are not. You consume deepfakes with your own eyes, making it understandably difficult to question their veracity. As such, laws are being proposed around the country regulating their creation. While sincere in their intent, legal penalties are insufficient. Deepfakes can be published from any jurisdiction anywhere in the world. Centralized forces cannot effectively counter a decentralized threat. 

That’s where blockchain has a unique advantage. Unlike almost all centralized entities, blockchain is trustless. Meaning that accepting the authenticity of data added to a block does not require belief in an individual or even an organization, it simply asks you to trust the underlying code –– which is typically open source and publicly auditable. The absence of human control makes it so you cannot, however hard you try or powerful you are, change what is permanently cemented on-chain. Governments attempting to label genuine information as AI-generated will be out of luck. As a public utility, blockchain is accessible to everyone but controlled by no one, creating an equal, decentralized playing field for information storage. 

That said, blockchain will not single-handedly eliminate deepfakes. A malicious actor could, for instance, fabricate a video of the U.S. president warning of imminent nuclear strike. Even if this contradicted the verified record of official statements coming from the White House, the mere shock of nuclear war could deceive people. Not to mention the fact that many news-worthy events solely rely on citizen-journalists to provide primary-source material. While much of this material is likely to be accurate, people could still manufacture deepfakes and pass them off as legitimate. 

Think of blockchain as a publically-accessible ledger where trusted sources can upload unalterable material. The technology itself does not verify truth, but when true content is uploaded on-chain, it becomes permanently incorruptible. So while blockchain is not restricted to well-meaning people, those people will be its primary beneficiaries. 

But for this to work, mass adoption is essential. A great place to start is for governments to begin recording key decisions and transactions on-chain, creating an immutable historical record resistant to AI manipulation. The same goes for news organizations, which could use blockchain-based decentralized storage to preserve and publish primary source material. At the individual level, uploading important documents (birth certificates or tax information, for example) on-chain will reduce the risk of AI-enabled identity fraud and data breaches. Imagine deepfakes of your passport. 

For all the damage deepfakes are currently doing to our information ecosystem, we are in the infant stages. The naked eye test still works to distinguish real from phony, but this won’t be the case for long. While it is necessary to spread awareness and concern about deepfakes, we should not let this destroy any hope of a healthy news landscape. Surrendering to a “Mountainhead” of AI-fueled anarchy is as reckless as it is preventable.

Jack Verrill is an Opinion Analyst writing about national security, the economy and domestic politics. He can be reached at [email protected].

Related articles