Last updated:
Reporter
Rachel Wolfson
Reporter
Rachel Wolfson
Share
Why Trust Cryptonews
Artificial intelligence (AI) is impacting every sector, yet it’s believed that AI will heavily disrupt the blockchain industry moving forward.
Recent market insights show that the intersection of AI and blockchain technology is expected to be worth over $2.7 billion by 2031.
Web3 Platforms Show Interest in AI Agents
While various use cases are still underway, a number of Web3 companies have started to incorporate AI agents to perform specific tasks.
Danny O’Brien, Senior Fellow at Filecoin Foundation, told Cryptonews that AI agents are autonomous software systems that leverage one or many AI models. O’Brien explained that these agents have access to data and tools, allowing them to run a sandbox, query databases or even execute a smart contract.
“Web3 developers are building agents that can handle complex tasks, while also paving the way for collectives of agents that can specialize in completing multi-step tasks,” O’Brien said.
Ron Bodkin, Co-Founder and CEO of Theoriq – an AI communication platform – added that unlike large language models (LLMs) such as ChatGPT, AI agents are designed to achieve specific goals or perform certain functions.
“This is often by processing selected input data, analyzing it, and taking appropriate actions based on predefined rules or learned behavior,” Bodkin said. “ For example, the Theoriq protocol is purpose-built to give agents planning, routing, self-reflection and iteration capabilities – via teams of agents collaborating in ‘Collectives.’”
AI Agents Make It Easier To Access Web3 Platform
O’Brien shared that Filecoin Foundation – the organization behind the decentralized data storage network Filecoin – will be using AI agents. Filecoin Foundation recently entered into a partnership with Theoriq to develop AI Agents trained using data hosted on the Filecoin network.
“These agents will harness artificial intelligence to make data more accessible, consumable, and usable for a variety of audiences, while highlighting the significance of transparent, verifiable sources,” O’Brien said.
In May, Filecoin Foundation demoed a Filecoin AI agent developed in collaboration with Theoriq. O’Brien mentioned that this agent was trained on the Filecoin docs and GitHub repositories, and allows users to get detailed answers through natural language queries.
“Users can use this agent to easily learn how to build on Filecoin, troubleshoot common issues, get started as a storage provider, and more,” he remarked.
This is important, as it’s been noted that building on Web3 platforms, along with accessing them, can often be a complex task.
Aanchal Malhotra, Head of RippleX Research – the research team behind blockchain-based digital payment network Ripple – told Cryptonews that AI agents are also being used by Ripple.
“AI chatbots are part of Ripple’s commitment to making blockchain development on the XRP Ledger more accessible and less time-consuming, enabling developers, especially those new to the field, to quickly receive answers to their queries, speeding up the process from concept to application,” Malhotra said.
In addition, the layer-2 network Rootstock developer portal now offers AI coding support.
Using AI Agents For Real World Use Cases
AI agents are also being used to solve real world problems. For example, O’Brien shared that Filecoin Foundation and Theoriq will continue working together to develop agents for a wide range of data hosted on Filecoin. This will provide seamless natural language interaction with public datasets.
“Filecoin Foundation and Theoriq are exploring a CIA AI agent that will enable researchers and policymakers to efficiently search and verify one million declassified documents,” he said. “This will allow users to leverage 25 years of declassified CIA datasets from MuckRock that are stored on Filecoin.”
O’Brien elaborated that this is important, as the AI agent will be able to parse millions of documents quickly – versus manually going through them. “Agents will be able to quickly provide natural language responses to inquiries about the data,” he said.
Challenges with AI Agents
Although AI agents are innovating within Web3, a number of challenges remain. O’Brien pointed out that majority of data from the web is currently gathered, stored, and provided by just a few powerful companies. Unfortunately, this silos information, centralizes control, and leaves data vulnerable to single points of failure.
O’Brien believes that the same may be true for AI. “We’re following the same path with AI, where AI developers are locking down data and code, which impacts the visibility into how AI models work, like what data they’re using and how they’re trained,” he said.
While this may be, O’Brien noted that platforms like Theoriq are taking a different approach by ensuring decentralized, open, and transparent AI.
Malhotra added that one of the greatest challenges facing the incorporation of AI agents into Web3 platforms is how to fairly compensate data providers and incentivize content creation for responsible AI development.
“The nature of decentralization and user data ownership conflicts with AI’s need for large datasets to train and operate effectively,” she said.
Malhotra pointed out that blockchain-based solutions to overcome this issue could include tokenizing AI training models as non-fungible tokens (NFTs).
“This would allow developers to earn royalties when their models are used to train AI agents,” she said. “There’s also privacy-preserving techniques like federated learning, where models can be trained on distributed data without the need to centralize it. Zero-knowledge proofs and homomorphic encryption also allow AI to process encrypted data without needing to access the underlying information.”
Another significant hurdle Malhotra mentioned is the so-called “black-box problem”. “There is a lack of visibility into how AI systems arrive at their conclusions or predictions, leading to concerns over authenticity, correctness, bias and fairness,” she remarked.
Web3 Companies Will Continue Incorporating AI Agents
Challenges aside, O’Brien believes that Web3 platforms will continue to use AI agents. “The possibilities are endless,” he said. “Instead of processing an overwhelming amount of data in real time, AI agents can help focus the power of AI technologies – thereby amplifying their impact for specific purposes.”
This seems to be the case. For example, Malhotra shared that a number of opportunities exist to harness AI to provide enhancements to the functionality of the XRP Ledger.
“Specific areas where AI could improve functionality on the XRPL in future include the automation of data-driven decision-making through real-time analysis of transaction patterns, network activity, and resource usage,” she said. “AI can also be used for the development of algorithmic trading strategies to enhance investment decisions, reduce human bias, and maximize profitability.”
