An article reviewing the directions and protocols of AI empowering Crypto

Author: Nan Zhi, Odaily Planet Daily

At the end of 2022, ChatGPT opened for public testing, sparking a wave of interest in LLM-type AI. Since then, the number of AI projects has rapidly increased, with various subcategories including data, AI models, computing power, applications, and more. As we enter 2024, what are the angles of integration between AI and Crypto applications? Has AI brought a productivity revolution to Crypto? Odaily Planet Daily will review various integrated applications in this article.

AI Auditing

As early as last year during the initial boom of LLMs, Slow Mist used ChatGPT for auditing tests. At that time, the AI was general-purpose, and ChatGPT was tested by Slow Mist to have the ability to check for ordinary small-scale code vulnerabilities, but it was insufficient for handling complex long code issues. Recently, several applications specifically aimed at Web3 code auditing have emerged:

AEGIS AI

AEGIS AI is designed specifically for Web3 code auditing, primarily including static code auditing and token auditing (i.e., auditing the open-source contract code of tokens). The auditing will be handled by three types of AI: Auditors for code auditing, Reviewers for rechecking and fact verification, and Judges for assessing risk levels.

Additionally, AEGIS AI provides real-time monitoring (WatchDog) to detect live activities and proactively identify suspicious behavior.

ZAN

ZAN is a plug-and-play tool and service aimed at projects. In addition to static analysis (i.e., basic vulnerability checks) and fuzz testing (using simulated transactions for testing), ZAN also offers AI-based auditing services, utilizing GPT fine-tuned on a large-scale vulnerability database for security checks.

AI Trading / Strategy

These projects use AI for trading assistance, aiming to provide users with trading signals and a better trading experience.

AlphaScan AI

AlphaScan AI connects thousands of Telegram groups, allowing users to manually set search signals, while AlphaScan AI conducts full-channel information searches and automated trading. It also supports advanced analysis of content using AI to push key information.

Hera Finance

Hera Finance is an AI-driven multi-chain DEX aggregator, aiming to provide the best trading pair Router using AI. Hera has launched an AI pathfinder, which has self-learning capabilities, with the fundamental goal of finding the maximum output solution for users' Swap operations (i.e., exchanging for as many tokens as possible).

Aimbot AI

Aimbot AI is a relatively unique project, where the team has developed an AI bot specifically for on-chain meme coin trading. Users can share in the trading profits of the AI by purchasing project tokens. As of the time of writing, Aimbot AI has made a profit of 716.14 ETH through on-chain token trading.

AI Search / Q&A

These products use AI to integrate, train, and output information from platforms like X, TG, and various project news, with fairly consistent functions. They typically include hot topic tracking and Web3 question answering, capable of providing good overviews of basic concepts, historical events, and hot topics. However, the current depth of understanding is still lacking, and they can only serve as references or basic materials, not as independent decision-making tools.

Due to strong product homogeneity, further analysis will not be provided here. Some existing products include Scopechat AI, QnA3, AwesomeQA, Brian, Libertai, Clab.AI, etc.

AI Assistants

These products use AI to enhance existing platform functionalities, lowering the usage threshold. Scopechat AI mentioned in the previous section also falls into this category, along with other projects including:

Nansen AI

Nansen AI's features include on-chain data anomalies, intelligent search, and filtering, grouping, and tracking specific wallets. The intelligent search (AI-powered Smart Search) allows users to ask questions in natural language, such as the number of holders or transaction volumes, and Nansen AI will search for data and provide answers on the site and related sites.

Dune AI

Dune AI allows users to ask questions in natural language, automatically generating SQL query code based on user input. It can also calculate the credibility difference between the generated query code and the user's expected values, supporting secondary editing of queries.

Conclusion

Looking back at the AI+Crypto track, there are more products that empower AI with Crypto, such as decentralized computing power and shared data. While there have been various explorations of AI empowering Crypto, the depth and breadth are still far from sufficient, and the market is still awaiting a disruptive product.