The solution to this problem lies in developing solutions that revolve around model benchmark testing. However, the challenges surrounding model benchmark testing, such as cost, throughput, and quality trade-offs, make positive competition in this field challenging. BitTensor is one of the largest cryptocurrencies focused on artificial intelligence and aims to address this problem. However, there are still some technical challenges that may hinder its widespread application (see Appendix 1).
Furthermore, trustless model inference, which proves that the model’s output is actually generated by the claimed model, is another area of positive research in Crypto x AI. However, we believe that these solutions may face challenges in terms of demand as the size of open-source models shrinks.
In a world where models can be downloaded and executed locally, and content integrity can be verified through established file hash/verification methods, the role of trustless inference becomes less clear. While many people still cannot train and execute LLMs on lightweight devices like smartphones, powerful desktop computers (such as those used for high-end gaming) can already execute many high-performance models.
Data Sources and Identity
As the output of generative AI becomes increasingly difficult to distinguish from human output, the importance of tracking AI-generated content has become a focus of concern. GPT-4, with its Turing test speed three times faster than GPT-3.5, will likely reach a point where it becomes indistinguishable whether online personalities are from machines or real humans. In such a world, determining the humanity of online users and watermarking AI-generated content will become crucial features.
Decentralized identifiers and personality verification mechanisms like Worldcoin aim to address the previous problem of identifying humans on-chain. Similarly, releasing data hashes to the blockchain can help verify the age and source of content, thereby empowering data sources. However, like the previous section, we believe the feasibility of Crypto-based solutions must be weighed against centralized alternatives.
Further research is underway regarding AI watermarking to embed hidden signals in text and image outputs, allowing algorithms to detect whether the content is generated by AI. Many leading AI companies, including Microsoft, Anthropic, and Amazon, have publicly committed to adding such watermarks to their generated content.
Additionally, for compliance reasons, many existing content providers have strict records of metadata retention. Therefore, users typically trust metadata associated with social media releases (although not their screenshots), even if they are stored centrally.
It should be noted that any Crypto-based data source and identity solution needs to integrate with user platforms to be widely effective. Thus, while Crypto-based solutions are technically feasible in proving identity and data source, their adoption is not a given and will ultimately depend on business, compliance, and regulatory requirements.
Trading AI Narratives
Despite the aforementioned issues, starting from Q4 2023, the performance of many AI tokens has outperformed Bitcoin, Ethereum, and major AI stocks such as Nvidia and Microsoft. We believe this is because AI tokens benefit from the broader Crypto market and the associated hype around artificial intelligence (see Appendix 2).
Therefore, even during a Bitcoin price decline, AI-focused tokens experience price volatility that may be upward during Bitcoin’s downturn. Figure 5 shows the performance of AI tokens on days when Bitcoin trading declined.
Overall, we still believe that the AI narrative trading lacks many short-term sustainable demand drivers. Due to the lack of clear adoption predictions and indicators, various meme-like speculative sentiments dominate the space, which in our view may not be sustainable in the long run.
Ultimately, price and utility will converge, and the unanswered question is how long it will take and whether utility will rise to match prices, or vice versa. In other words, we do believe that a sustainable and constructive Crypto market and strong Crypto AI narrative may be maintained for a period of time.
Conclusion
The role of Crypto in AI does not exist in a vacuum, as any decentralized platform competes with existing centralized alternatives and must be analyzed against broader business and regulatory requirements. Therefore, we believe that simply replacing centralized providers with “decentralized” ones is not enough to drive meaningful market adoption. Generative AI models have existed for several years and have maintained a certain degree of decentralization due to market competition and open-source software.
A recurring theme in this report is that while Crypto-based solutions are technically feasible, they require significant work to achieve functional parity with more centralized platforms, assuming these platforms do not stand still during this period. In fact, due to consensus mechanisms, centralized development often outpaces decentralization, which can pose challenges to the rapidly evolving field of artificial intelligence.
Given this, we believe that the overlap of AI and Crypto is still in its early stages and may change rapidly in the next few years with the development of the broader AI field. The decentralized AI future envisioned by many within the Crypto community is currently uncertain, and the future of the AI industry itself is largely undecided. Therefore, we believe it is wise to proceed cautiously in such a market and conduct deeper research on how Crypto-based solutions can truly offer meaningful and better alternatives or at least understand the underlying narrative of transactions.