The emergence of DeepSeek has shifted the understanding of what AI can accomplish on a comparatively small budget. However, as groundbreaking as the model is, DeepSeek still suffers from many of the limitations that have plagued other AI models, including the reliability of data inputs and the transparency of information.
As Joel Hugentobler, Cryptocurrency Analyst at Javelin Strategy & Research, found in his Harnessing AI Through Blockchain report, blockchain can be not only the solution for these issues but also the best foundation for one of the most powerful technologies in recent times.
Escaping the Black Box
One of the main issues with AI is that it can provide false or misleading information. This is a problem resulting from centralization—AI is making decisions based on a repository of knowledge that has discrete boundaries.
Another concern is that data scientists often don’t have full transparency into what artificial intelligence is up to within these parameters. This has led to the “black box” problem, where AI has made the wrong decision, but analysts can’t understand why. This issue is exacerbated when AI is faced with a substantial number of variables, as can occur in making complex financial decisions.
A decentralized foundation like blockchain can mitigate both issues. Blockchain is transparent and its records are immutable, so scientists can get full clarity into the data inputs feeding the model and decisions at every step.
AI models can be even more efficient when they are open-source because there is a decentralized community that can ensure the model is optimized and on track. Decentralized AI also distributes tasks that are normally centralized in large data centers across the network.
“Open source, especially paired with blockchain, is the trend going forward,” Hugentobler said. “It’s more efficient, and it eliminates a single point of failure. Moving away from a typical AI model running if-then logic to a more dynamic approach integrating blockchain propels both technologies forward and enables companies to use it for more things.”
Dynamic Smart Contracts
Some of the most dynamic efficiencies gained from shifting AI to the blockchain come from supercharged smart contracts. Smart contracts are digital contracts on the blockchain that execute when certain conditions or thresholds are met.
This could include tasks like issuing a ticket, selling a stock, or sending out a push notification. Once the smart contract executes, the blockchain is updated, it can’t be changed, and the pertinent parties can immediately view the results.
Smart contracts can also be stacked to automate a workflow, with a sequence of actions that are completed in a domino effect. However, when AI and blockchain are combined, smart contracts have the potential to do much more.
“With normal AI, it’s like if Apple stock reaches $60, then sell,” Hugentobler said. “With the dynamic approach with blockchain and AI, it’s if Apple stock is expected to rise within a couple of months based on sentiment or volume, then hold. If Apple stock breaks above the 52-week high on more than average volume, then buy. Otherwise, if it breaks $40 on more than average volume, sell.
“All that can be embedded in the smart contract, so it happens automatically. It can be applied to the stock market, compliance, know your customer (KYC), you name it. In that dynamic model, rather than just the if-then approach, it opens the door to more automation.”
Decentralizing Privacy and Security
Substantial buzz has swirled around the potential for AI in many use cases, but it has been somewhat mitigated as the limits of artificial intelligence have been exposed. In addition to the incidents where AI has provided bad information, there are also privacy and security concerns.
For example, DeepSeek has already been banned from government devices in many countries—including the United States—over concerns about the model’s ties to the Chinese Communist Party (CCP) and the lack of transparency about how DeepSeek uses its data.
In a letter, U.S. lawmakers said that “by using DeepSeek, users are unknowingly sharing highly sensitive, proprietary information with the CCP—such as contracts, documents, and financial records.”
These privacy concerns have also been raised about other centralized AI models because they collate vast amounts of data, often without user consent. There have also been many instances where the data in AI systems has been tampered with, either to spread misinformation or to perpetrate criminal acts.
Blockchain is a better solution because its unchangeable records are fully secure, which drastically reduces the risk of bias or manipulation. The decentralized approach also means users retain control over the data they share with AI.
Integrating Both Technologies
The benefits of digital asset technologies have caused record-high investments by the leading financial institutions over the past few years. Tokenization, stablecoins, and crypto have become far more prevalent—and they are all underpinned by blockchain technology.
Even though AI and digital assets have emerged in disparate arenas, the synergies these technologies share mean they could be better together.
“The crypto industry has been around for 15 to 16 years now, and it’s grown to what it has become today without the help of AI,” Hugentobler said. “But integrating both of those technologies is just going to speed up the pace of change and evolution. I think that’s going to spill into a lot of other areas rather quickly.
“Financial institutions need to assess their use of AI, and they also need to assess their infrastructure. If they can integrate blockchain into their AI systems, there’s a lot that they can do that agentic AI really can’t do.”
