Grayscale: In the AI era, how can encryption be used to its full potential?

Original author: Will Ogden Moore

Original text compilation: Luffy, Foresight News

Related Reading:

“AI wave strikes again, an article lists Grayscale AI Fund Holdings projects”

Artificial Intelligence (AI) is one of the most promising emerging technologies of this century, with the potential to exponentially increase human productivity and drive medical breakthroughs. Although AI has already made its mark, its future impact will only grow. PwC estimates that by 2030, it will become a massive industry worth $15 trillion.

However, this promising technology also faces challenges. As artificial intelligence technology becomes more powerful, the AI industry becomes highly centralized, with power concentrated in the hands of a few companies, which poses a potential threat to the entire human society. AI has also raised serious concerns about the falsification, bias, and data privacy risks in Depth. Fortunately, Crypto and its Decentralization and transparent characteristics offer potential solutions to some of these problems.

Below, we will explore the problems caused by centralization and how Decentralization artificial intelligence can help solve some of them, and discuss the intersection of Cryptocurrency and artificial intelligence, with a focus on encryption applications that have shown early adoption in this field.

Problems with Centralized Artificial Intelligence

Nowadays, the development of artificial intelligence is facing certain challenges and risks. The network effect and intensive capital requirements of artificial intelligence are very significant, so that artificial intelligence developers outside of large technology companies, such as small companies or academic researchers, either have difficulty obtaining the resources needed for development or cannot commercialize them. This limits the overall competition and innovation of artificial intelligence.

As a result, the influence of this key technology is mainly concentrated in the hands of a few companies such as OpenAI and Google, which has raised serious concerns about the governance of artificial intelligence. For example, in February of this year, Google’s artificial intelligence image generator Gemini exposed racial bias and historical errors. In addition, in November last year, a six-member board decided to dismiss OpenAI CEO Sam Altman, exposing the fact that a few people control these companies.

With the increasing influence and importance of artificial intelligence, many people are worried that a company may have the decision-making power over AI models that have a huge impact on society, and it may set barriers, operate behind closed doors, or manipulate the models for its own gain.

How Decentralization Can Help Artificial Intelligence

Decentralized AI refers to the use of blockchain technology to distribute ownership and governance of AI in a way that improves transparency and accessibility. Grayscale Research believes that Decentralized AI has the potential to liberate these important decisions from closed systems and hand them over to the public.

Blockchain technology can help developers have longer access to artificial intelligence, dropping the barriers for independent developers to develop and commercialize. We believe this can help improve innovation and competition in the artificial intelligence industry, achieving some kind of balance between small companies and tech giants.

In addition, Decentralization AI helps to achieve the democratization of AI investment. Currently, there are few ways to obtain financial benefits related to AI development, except for a few technology stocks. At the same time, a large amount of capital from private sales is allocated to AI startups and private companies (47 billion USD in 2022 and 42 billion USD in 2023). Therefore, only a small number of venture capitalists and accredited investors can obtain financial benefits from these companies. In contrast, Decentralization’s AI encryption assets are equal for everyone, and everyone can own a part of AI’s future.

How far has this interdisciplinary field developed?

The intersection of Cryptocurrency and artificial intelligence is still in its early stages, but the market response is encouraging. As of May 2024, the AI concept encryption assets (defined by Grayscale Research as a Cryptocurrency portfolio including NEAR, FET, RNDR, FIL, TAO, THETA, AKT, AGIX, WLD, AIOZ, TFUEL, GLM, PRIME, OCEAN, ARKM, and LTP) have a return rate of 20%, second only to currency concept category (Figure 1). In addition, according to data provider Kaito, artificial intelligence is currently the hottest ‘narrative’ on social platforms compared to other topics such as Decentralized Finance, Layer 2, Memecoin, and real-world assets.

Recently, some well-known figures have begun to embrace this emerging intersection, dedicated to addressing the shortcomings of centralized artificial intelligence. In March of this year, Emad Mostaque, the founder of the well-known artificial intelligence company Stability AI, left the company to explore Decentralization AI, stating, “Now is the time for AI to open up and embrace Decentralization.” Cryptocurrency entrepreneur Erik Vorhees recently launched Venice.ai, a privacy-focused AI service with end-to-end encryption capabilities.

Figure 1: So far this year, the performance of the AI field has been almost better than all sub-sectors of Cryptocurrency

We can divide the integration of Cryptocurrency and artificial intelligence into three main subcategories:

• Infrastructure Layer: The network that provides platforms for AI development (such as NEAR, TAO, FET);
• Resources needed for artificial intelligence: provide key resources such as computing, storage, data, etc. required for AI development (e.g., RNDR, AKT, LPT, FIL, AR, MASA);
• Solve AI problems: attempting to address AI-related issues, such as the rise of robots and Depth forgery, as well as model validation (e.g., WLD, TRAC, NUM).

Figure 2: Project map of the fusion of artificial intelligence and encryption, source: Grayscale Investments

AI Infrastructure Network

The first type is to provide a network with an open architecture that does not require permission, built specifically for AI development. These networks do not focus on a particular AI product or service, but rather create underlying infrastructure and incentive mechanisms for various AI applications.

NEAR stands out in this category, with one of the co-founders being a co-author of the ‘Transformer’ architecture that powers AI systems like ChatGPT. However, the company recently leveraged its AI expertise to announce the development of ‘user-owned AI’ through a research and development department led by a former OpenAI research engineer. In late June 2024, Near launched an AI incubator program for developing native models, AI application data platforms, AI agent frameworks, and computing markets.

Bittensor is another notable example. Bittensor is a platform that uses TAO Token to economically incentivize the development of artificial intelligence. Bittensor is the underlying platform of 38 subnets, each with different use cases, such as chatbots, image generation, financial forecasting, language translation, model training, storage, and computation. The Bittensor network rewards the best performing miners and validators in each subnet with TAO Token and provides developers with an unlicensed API to help them build specific artificial intelligence applications.

The AI infrastructure network also includes other protocols, such as Fetch.ai and Allora. Fetch.ai is a platform for developers to create complex AI agents (i.e. ‘AI agents’), recently merged with AGIX and OCEAN, with a total value of approximately $7.5 billion. Another one is the Allora network, which focuses on applying AI to the financial sector, including automatic trading strategies for decentralized exchanges and prediction markets. Allora has not yet launched its token and conducted a strategic financing round in June, raising a total of $35 million.

Provide the resources needed for AI

The second type is projects that provide resources needed for artificial intelligence development in the form of computation, storage, or data.

The rise of artificial intelligence has created an unprecedented demand for GPU computing resources. Decentralization GPU markets such as Render (RNDR), Akash (AKT), and Livepeer (LPT) provide idle GPU supplies for developers needing computation for model training, model inference, or rendering 3D generative AI. Render is estimated to offer around 10,000 GPUs, with a focus on artists and generative AI, while Akash offers 400 GPUs, targeting AI developers and researchers. Meanwhile, Livepeer has recently announced its new AI subnet plan, aiming to complete text-to-image, text-to-video, and image-to-video functions by August 2024.

In addition to requiring a large amount of computation, AI models also require a large amount of data. Therefore, there is a significant increase in the demand for data storage. Data storage solutions such as FIL (FIL) and Arweave (AR) can be used as alternatives to storing AI data on centralized AWS servers. These solutions not only provide economically efficient and scalable storage, but also enhance data security and integrity by eliminating single points of failure and reducing the risk of data leakage.

Finally, existing AI services such as OpenAI and Gemini continuously access real-time data through Bing and Google searches. This puts all other AI model developers outside of the tech giants at a disadvantage. However, data scraping services like Grass and Masa (MASA) can help create a fair competitive environment by allowing individuals to commercialize their application data for AI model training while maintaining control and privacy over their personal data.

Solve AI related issues

One major issue exacerbated by artificial intelligence is the proliferation of robots and fake information. The spate of Depth-generated fake content has already influenced presidential elections in India and Europe, and experts are ‘very afraid’ that the upcoming presidential elections will be engulfed in a ‘tsunami of fake information’ driven by Depth forgeries. Projects aimed at addressing the issue of Depth forgery by establishing verifiable sources of content include Origin Trail (TRAC), Numbers Protocol (NUM), and Story Protocol. Additionally, Worldcoin (WLD) seeks to address the robot issue by proving a person’s humanity through unique biometric technology.

Another risk of artificial intelligence is ensuring trust in the model itself. How do we trust that the AI results we receive have not been tampered with or manipulated? Currently, several protocols are working to address this issue through cryptography, Zero-Knowledge Proof, and fully Homomorphic Encryption (FHE), including Modulus Labs and Zama.

Conclusion

While these Decentralization artificial intelligence assets have made initial progress, we are still in the early stages of this intersection. Earlier this year, renowned venture capitalist Fred Wilson said that artificial intelligence and Cryptocurrency are “two sides of the same coin,” and “Web3 will help us trust artificial intelligence.” As the artificial intelligence industry continues to mature, Grayscale Research believes that these encryption use cases related to artificial intelligence will become increasingly important, and these two rapidly developing technologies may support and develop together.

Many signs indicate that the era of artificial intelligence is coming, which will have profound impacts, both positive and negative. By leveraging the characteristics of blockchain technology, we believe that Crypto can ultimately help mitigate some of the dangers brought by artificial intelligence.

Original Link