Of all the fears running rampant over Artificial Intelligence (AI) these days, perhaps the most significant one is the rise of deep fakes. The technology to create false text, images, and even audio and video has evolved to the point where only fine-grained analysis can spot the frauds.
But while deep fakes are usually spoken of in the context of mass deception – false statements from celebrities or political leaders, for example – their penchant for deceiving individuals is also a major cause for concern. Frantic telephone calls from grandchildren or siblings in desperate need of money or video messages from tax authorities threatening legal action are only a few of the ways in which deep fakes can bilk even tech-savvy citizens.
AI not only allows tricksters to create highly believable fakes, but it also helps disseminate them to vulnerable targets as well. It can optimize the messaging to draw in key demographics and help shield itself against detection from even the most advanced digital forensic technologies.
It isn't just AI generated text that is starting to bleed over into search results.
The main image if you do a Google search for Hawaiian singer Israel Kamakawiwoʻole (whose version of Somewhere Over the Rainbow you have probably hear) is a Midjourney creation right from Reddit. pic.twitter.com/VdZmuUAwCK
— Ethan Mollick (@emollick) November 27, 2023
This is why many people consider blockchain an effective weapon against deep fakes. Distributed ledger technology’s primary benefit is its immutability: once a record has been entered into the chain, it is virtually impossible to alter it. This was initially intended to ensure the sanctity of financial transactions, but it can also be used to establish the provenance of any digitized asset, including text, audio, and video.
Blockchain-based tools to verify currency transfers, real estate transactions, and even products in supply chains have been around for a while. However, only lately has research shifted to affirming the authenticity of digital media content.
Groups like the Coalition for Content Provenance and Authenticity (C2PA) have harnessed the efforts of Microsoft, Adobe, and Intel to devise new tools that track timestamps, transaction hashes, and other data to determine ownership of individual pieces of digital content.
These are primarily intended to help media organizations protect their content, but there are no consumer-grade blockchain solutions to protect people from swindlers.
Tracking Through Time
One of the more common tools used to unveil deep fakes is the digital signature. These artifacts can be embedded into original content to establish ownership, but once that content is released into the digital universe, digital signatures suffer from two main flaws: distribution and anchoring the signature in time. Blockchain solves both problems.
Ideally, a digital signature can help determine if a copy is authorized or not, but it doesn’t always work this way in practice. Simple copying and pasting or changing the work into a different format is enough to erase a signature. However, if it is embedded in blockchain, ownership can be traced back to the original, and any stolen or altered versions can be easily identified. At the same time, blockchain helps preserve the identity of the creator since that information is hidden from view.
Blockchain also brings a number of other tools to help thwart deep fakers. Decentralized validation requires multiple chain nodes to sign off on any transaction, making it extremely difficult to alter or manipulate content without a trace.
Most chains are also empowered with content ownership and rights management tools to protect intellectual property rights. This has the twin effect of ensuring owners are compensated properly for their original creations and preventing unauthorized use.
We can expect these tools to become more powerful over time now that AI models like machine learning and intelligent analytics are entering the development pipelines of today’s most popular blockchain platforms.
In the near future, content creators will have a broadened capacity to continuously monitor social media and other outlets for faked versions of their content and stifle their dissemination.
Gaps in the Armor
One thing to keep in mind is that while blockchain’s immutability may provide an effective weapon against deep fakes, blockchain platforms themselves are not immune to security breaches.
Patrick J. Simon, president and manager of Beehive Technology Solutions LLC, notes that blockchain operates on multiple protocols, which means they require any number of integration layers to work together. Any time there is an interface between two systems, there are security vulnerabilities.
Exploiting those vulnerabilities is possible in a number of ways, but the primary means is key theft. If hackers gain access to a legitimate account, they can launch faked content as original material. And just like any other data environment, blockchains themselves are only as safe as their security and data protection measures allow, and these tools are coming under increasingly sophisticated assault with the advent of AI and quantum computing.
This goes to the very essence of blockchain’s ability to fight deep fakes: it cannot separate fact from fiction, it can only provide a means to quickly determine if a piece of content is a faithful copy of its original.
Once that has been established, the actual truth of its message is up for debate.