Web3 companies deploy task-performing AI agents to assist improve platform accessibility for developers and users.
Although artificial intelligence (AI) is having an impact on every business, it is anticipated that AI will significantly disrupt the blockchain sector in the near future.
According to recent industry research, the combined value of blockchain technology plus artificial intelligence (AI) is projected to reach over $2.7 billion by 2031.
AI Agents Are of Interest to Web3 Platforms
While numerous use cases are still being developed, several Web3 businesses have begun implementing AI agents to carry out particular duties.
According to Danny O’Brien, Senior Fellow at the Filecoin Foundation, AI agents are self-governing software programs that make use of one or more AI models, as reported by Cryptonews. According to O’Brien, these agents can operate a sandbox, query databases, or even carry out a smart contract because they have access to data and tools.
Web3 developers, according to O’Brien, “are creating agents that can manage complicated tasks while also laying the groundwork for collectives of agents that can specialize in finishing multi-step tasks.”
The CEO and co-founder of Theoriq, an AI communication platform, Ron Bodkin, noted that AI agents are made to accomplish particular tasks, as opposed to large language models (LLMs) like ChatGPT.
According to Bodkin, “this is often accomplished by processing specific input data, analyzing it, and acting appropriately based on predefined rules or learned behavior.” For instance, the Theoriq protocol was designed with the express intention of enabling agents to engage in self-reflection, planning, routing, and iteration through the formation of “Collectives” of cooperating agents.
AI Agents Simplify Web3 Platform Access
According to O’Brien, AI agents will be used by the Filecoin Foundation, the company that created the decentralized Filecoin data storage network. Theoriq and the Filecoin Foundation have partnered to create AI agents that are trained on Filecoin network data.
Utilizing AI Agents in Practical Applications
Real-world issues are also being resolved by AI bots. O’Brien revealed, for instance, that Theoriq and the Filecoin Foundation will keep collaborating to create agents for a variety of data held on Filecoin. Natural language engagement with public datasets will be seamless as a result.
He stated, “Theoriq and the Filecoin Foundation are investigating a CIA AI agent that will allow scholars and decision-makers to effectively search and validate one million declassified documents.” Users will be able to take advantage of 25 years’ worth of MuckRock declassified CIA datasets that are kept on Filecoin thanks to this.
According to O’Brien, this is significant because, as opposed to wading through millions of papers by hand, the AI agent will be able to digest them fast. “When asked questions about the data, agents will be able to respond in natural language quickly,” he stated.
Difficulties with Artificial Intelligence
Despite the fact that AI agents are advancing Web3, certain obstacles still need to be overcome. O’Brien brought out the fact that a small number of very strong businesses currently collect, store, and distribute the majority of the data found on the internet. Regrettably, this centralizes control, separates information, and exposes data to single points of failure.
O’Brien thinks AI might work in a similar way. He claimed that “we’re following the same path with AI,” meaning that programmers are limiting access to data and code, making it harder to see how AI models operate and what data they use and how they’re taught.
Though this may be the case, O’Brien pointed out that platforms like as Theoriq are adopting an alternative strategy by guaranteeing AI that is decentralized, transparent, and open.
Malhotra went on to say that paying data providers properly and encouraging content creation for ethical AI development are two of the biggest obstacles to the integration of AI agents into Web3 platforms.
“AI requires large datasets to train and function effectively, and the nature of decentralization and user data ownership conflicts with this,” the speaker added.
Malhotra noted that tokenizing AI training models as non-fungible tokens (NFTs) is one potential blockchain-based approach to solving this problem.
When their models are utilized to train AI agents, “this would allow developers to earn royalties,” the developer stated. “Other privacy-preserving methods include federated learning, which allows models to be trained on dispersed data without requiring centralization. AI can decrypt and process encrypted data without having to see the underlying data thanks to homomorphic encryption and zero-knowledge proofs.
Malhotra also noted the “black-box problem” as a major obstacle. “There are questions about authenticity, correctness, bias, and fairness because it is difficult to see how AI systems arrive at their conclusions or predictions,” she said.
Web-3 Businesses Will Keep Using AI Agents
Notwithstanding the difficulties, O’Brien thinks AI agents will be used in Web3 systems going forward. “There are countless possibilities,” he declared. “AI agents can assist in concentrating the power of AI technologies, thereby amplifying their impact for specific purposes, as opposed to processing an overwhelming amount of data in real time.”
Apparently, this is the case. Malhotra, for instance, mentioned that there are several chances to use AI to improve the XRP Ledger’s functionality.
The automation of data-driven decision-making through real-time analysis of transaction trends, network activity, and resource utilization are some specific areas where AI could improve functionality on the XRPL in the future, the speaker stated. “Algorithmic trading strategies can be developed using AI to improve investment decisions, minimize human bias, and maximize profitability.”