Blockchain fixes this: Building trust and privacy with AI | Opinion

Once thought of as distinct trends in technology—artificial intelligence and blockchain—are rapidly converging. These diverse sub-sectors include everything from decentralized energy and decentralized physical infrastructure networks to model fine-tuning and computation infrastructure.

Some skeptics might balk at startups that advertise their use of AI and blockchain, the two biggest buzzwords in business today, believing that they are simply trying to ride the hype of both. After all, if some companies already engage in AI washing, why not go a step further and also add blockchain to the mix? 

In truth, the intersection of AI and blockchain is not marketing hype. As I wrote in the H1 2024 report at HashKey Capital, AI and blockchain are uniquely complementary, enabling solutions far greater than the sum of their parts. 

Blockchain fixes the privacy concerns of AI 

Just imagine how much data you input into ChatGPT and other similar tools daily. Now multiply that by all the daily active users, along with the LLM developers who are ingesting terabytes of data for model training. In short, the world is entrusting a staggering amount of data to AI. 

GenAI thus represents a major attack vector for data breaches. For example, hackers could blackmail enterprises by releasing their employees’ GenAI logs, similar to how they threaten the leak of private data in ransomware attacks. These logs are especially sensitive given the breadth and depth of questions, content, data, and confidential information we input into our prompts.

Fortunately, one of the key tenets of blockchain is privacy: Through decentralization, users are spared the public scrutiny that often accompanies centralized systems, where they are subject to vetting by intermediaries and other large players. 

To this end, blockchain has developed several technologies that can enhance privacy, such as ZKML, OPML, and TEEML. While each presents its own trade-offs, their implementation will go a long way toward improving both data privacy and model privacy. By strengthening the security around AI, blockchain not only reduces the threat of data breaches but also fosters an all-important sense of trust in users. 

Blockchain makes AI sustainable

The rise of genAI also requires a vast amount of computational power, which strains the world’s energy. These demands will also increase exponentially as more individuals and businesses adopt AI, its use cases become more diverse, and it itself becomes more powerful, such as in its development of multimodal capabilities. 

These demands put companies in an ethical dilemma: While enterprises naturally want the enhanced productivity and efficiency that come with AI, they may rethink their use if it comes at the cost of harming the environment. This mindset is not only the right thing to do but also a matter of good business: 77% of consumers prefer patronizing environmentally and socially responsible companies. 

Blockchain can address the enormous energy demands of genAI through decentralized energy, which seeks to intelligently produce and distribute energy locally, a prospect more sustainable and efficient than centralized grids. With blockchain, companies do not have to choose between efficiency and environmental responsibility. They can envision a world improved by new innovations and progress toward that future sustainably. 

Blockchain may open AI

One criticism of AI is that its greatest benefits will be concentrated in the hands of the “few.” The minority that will most benefit from AI are model developers like OpenAI and Anthropic and large tech giants like Meta, Amazon, Google, and Apple, which have the most data. 

These critics argue that AI will enrich these players at the expense of everyone else. Developers will profit directly from their models, even though they were trained on datasets that included public data. Tech giants will be able to build the most profitable use cases around AI because they possess the most data about consumers.

This future is not ideal: AI should not be like gold or oil, where only a privileged elite reap the most benefits. Fortunately, blockchain may offer a solution here, too. Because blockchain is a public ledger, organizations can use it to share information more easily, including the data and even models central to AI.

Some companies have taken advantage of this propensity for information-sharing powered through the blockchain. One such company is Carv, a modular data layer that enables gaming and AI companies to better own, control, and monetize their data. Blockchain-powered solutions like Carv will create inclusion, extending the benefits of AI to individuals and organizations who may have otherwise been left out of this revolution.

The world’s biggest problems 

Blockchain was once criticized as a solution in search of a problem. With the rise of AI, blockchain will no longer be a niche technology on the margins of society. Because blockchain addresses some of the challenges with AI—namely, its security risks, energy demands, and potential unfairness—companies forward-thinking enough to employ both will enjoy competitive advantages in the marketplace, and the world will benefit. 

Businesses will be able to tackle the world’s biggest problems while ensuring data and model privacy. Despite the computational demands of this work, they can improve sustainability through the use of decentralized energy. Finally, and most of all, information-sharing over the blockchain can extend the technology to more people, ultimately enabling a more equitable distribution of AI’s transformative power. And that’s an imperative that we all must take seriously to unlock AI’s full potential. 

Jeffrey Hu

Jeffrey Hu is the head of investment research at HashKey Capital, where he monitors the trends and data moving the industry forward. He is also a regular contributor to the Bitcoin Optech newsletter to make Bitcoin’s key developments more accessible to the Chinese-speaking community.