Blockchain: a tool against deepfakes and misinformation?

Photo by Michał Kubalczyk on Unsplash  

Developments in artificial intelligence are transforming digital landscapes. While generative AI has its benefits, it also has negative repercussions. It is easier to create and disseminate hyper-realistic fake content (deepfakes) and misinformation. As it  becomes increasingly difficult to distinguish authentic content from forgeries, blockchain technology is emerging as a solution, specifically through its application in content provenance. 

Blockchain-based content provenance involves creating an unalterable record of a digital asset's origin and history on a distributed ledger. It acts as both a digital birth certificate and logbook. When a piece of content (a photo, video, audio recording or article) is created, a unique ‘fingerprint’ (a hash) of that content is generated and recorded on the blockchain. This record typically contains information such as: 

• Creator Identification: Verifying the original author or source. 

• Timestamp: A precise record of when the content was created. 

• Metadata: Details about the content (such as camera settings and location). 

• Modification History: Tracking any subsequent edits or alterations, with each change creating a new, linked entry on the blockchain. 

The decentralised nature of blockchain means that it is virtually impossible to tamper with the recorded information without the consensus of the entire network. This is the bedrock of trust in blockchain-based provenance. 

The development of generative AI has made it easier to create convincing content. Deepfakes in particular pose a threat to public trust, potentially being used for  propaganda, manipulating financial markets, or damaging reputations. Blockchain-based content provenance’s mechanisms can combat these threats. By creating an immutable record of a piece of content's origins, people can verify whether the content came from the original source. If a manipulated version surfaces, its blockchain record (or lack thereof) will immediately raise red flags. Secondly, if a piece of content with a blockchain record is altered without a corresponding entry on the chain, the discrepancy can be easily detected. This means that users can identify manipulated or fabricated content. Additionally, when content creators adopt blockchain-based provenance, they show a commitment to transparency and authenticity. This can help build trust with audiences and create a verifiable chain of custody for their work. Lastly, blockchain-based provenance puts the power of verification directly into the hands of consumers and users. Instead of relying solely on centralised fact-checking organisations (which can be slow or perceived as biased), people can independently verify the origin and integrity of the content they encounter online. 

Although not widespread, blockchain-based content provenance already has real-world applications. News organisations are exploring its use to combat the spread of fake news bearing their branding. Multimedia platforms are integrating blockchain to help artists protect their intellectual property and verify the authenticity of their work. However, there are challenges to widespread adoption. Technical complexities and interoperability remain key hurdles. User education and the development of intuitive tools for verification are also essential for mainstream acceptance. 

Despite these challenges, blockchain-based content provenance holds immense potential to address misinformation and deepfakes. As generative AI continues to evolve and the lines between reality and fabrication become increasingly blurred, the unforgeable fingerprint offered by blockchain could become an indispensable tool for safeguarding truth and trust in the digital age.