A report from financial information and analytics service S&P Global has indicated that blockchain technology can help with several risks associated with the use of artificial intelligence (AI). Titled “Crypto and AI: Shaping the Future of the Internet,” the S&P Global report states that through blockchain, the concerns surrounding widespread adoption of AI can be significantly mitigated.
Risks mentioned in the report include privacy, censorship, accountability, and data traceability. Although the report highlights these concerns, S&P Global believes that AI has perks that will benefit users in the long term.
The applications of blockchain technology are easily recognizable in the cryptocurrency sector. As one of the most popular applications of blockchain technology, crypto is commonly used as a medium of exchange for payments, and can easily simplify cross-border transactions. In addition, investors diversify their portfolios to include unique digital assets that are likely to spike. Users may find upcoming meme coins with unique advantages that are expected to surge as their level of adoption increases. This adoption is usually driven by exciting benefits like unmatched speed, decentralized governance for community involvement, and rewards obtainable from staking and yield farming.
Blockchain’s immutability
While cryptocurrency is blockchain’s most widely recognized application, S&P Global reports that this perception is gradually changing. According to the report, several entities have recognized that the technology can be used to address a wide range of concerns and issues. This is mostly due to blockchain’s immutability, as the ledger keeps a permanent and time-stamped record of all transactions it has processed. This record is verifiable right from the first transaction.
The immutability can help AI models to maintain transparency and compliance. Using blockchain, anyone can easily audit AI models and the information used to train them, especially if the blockchain is a scalable network that has a high throughput and charges minimal transaction fees. This transparency will ensure that the accountability concern is solved. If there is a public record that can be authenticated, regulators and lawmakers would be more willing to create policies supporting AI development and adoption because compliance is easily enforced.
Confirming provenance of AI training data
Since AI models are trained on data, provenance is important. It is essential to confirm where the data a model used came from and whether or not it is accurate.
In addition to accuracy, authorization is also a concern. Since AI usage exploded, several entities have sued ChatGPT creator OpenAI for using their creations without permission. Last year, a class action lawsuit filed in the United States District Court for the Northern District of California accused OpenAI and Microsoft of using illegal web scraping practices to gather data used to train its models. The lawsuit named several products, including ChatGPT-3.5, ChatGPT-4.0, Vall-E, and Dall-E.
Another issue is that AI models also tend to “hallucinate,” which is when they give false, misleading, or otherwise incorrect answers. Getting to the root of the hallucination and solving the problem requires verifying specific data inputs and ensuring that the data was not tampered with in any way. Blockchain technology can be used to confirm these inputs and ensure the data remains unchanged.
Privacy and censorship
Privacy is one of the biggest concerns with widespread AI use. Users and regulatory bodies are worried about the data shared with AI and what it is used for. Every day, users of popular AI tools like ChatGPT, Gemini, and Meta AI, input thousands of prompts that may contain personal or sensitive details depending on the request expected of the tool or chatbot. There are also AI virtual assistants that are fed with sensitive streams of data to carry out specific duties. While all of these are beneficial to the users, the endpoint of the data shared is a concern.
S&P Global believes that using blockchain technology to decentralize data control would be a sure way to solve this problem. For instance, several tools can be deployed to monitor information collection and usage, eventually triggering smart contracts to take preventive action or completely stop suspicious transactions or transfers.
Through blockchain technology, preventing censorship might be much easier than thought. Since decentralized networks are already distributed across hundreds or thousands of nodes, blockchain’s basic architecture is already resistant to censorship. Therefore, any content tagged wrong could be independently verified by tracing the source of its training and scrutinizing the disputed data. This makes it easy for organizations and corporate bodies to integrate AI into their processes without worrying about external censorship.
Featured image credit: Freepik
