Opinion by: Merav Ozair, PhD
Tech moguls cannot stop heralding the artificial intelligence revolution — from Bill Gates to Sundar Pichai to Jensen Huang — signaling that agentic AI and robotics will claim our jobs and act as our autonomous assistants performing on our behalf in our professional and personal lives.
Whether these scenarios happen in a few years or are decades away, we will most likely evolve into that future in some manner, and technology, once again, will reshape our lives. Without the support of blockchain technology, however, it would be quite difficult, and potentially impossible, for agentic AI and robotics to evolve to what its proponents expect them to.
If we expect these services and devices to act autonomously, security, privacy, transparency and accountability will be at the top of our minds. These areas are where blockchain shines and can support AI weaknesses to facilitate the scaling and evolution of this vision.
Blockchain strengths support AI weaknesses
Blockchain technology can significantly bolster the security of AI models by leveraging its key features such as decentralization, immutability, traceability, smart contracts, data privacy and identity verification. For example, but not limited to:
-
The decentralization aspect eliminates a single point of attack, increasing the resilience of AI models against breaches.
-
The immutability of blockchain ensures that the data used in training AI models and the models themselves cannot be illicitly altered, maintaining the integrity of the models.
-
Every alteration or decision made by the AI model can be audibly traced through blockchain, providing unparalleled transparency and accountability.
-
Smart contracts automate the enforcement of data access and usage rules, preventing unauthorized or unethical use of AI models.
-
Smart contracts can ensure that data is only used for training and testing and by authorized personnel, locking the option to be used for other purposes. Combining these rules with multiparty computation could prevent or at least mitigate AI adversarial attacks.
-
Blockchain allows secure multiparty computation, ensuring data privacy during AI model training by keeping the data decentralized.
-
Blockchain’s secure identity verification enhances the safety of AI systems by preventing unauthorized access.
Integrating AI with blockchain can establish a secure, transparent, traceable and decentralized AI environment, protecting our privacy, enhancing accountability and manifesting responsible AI.
Transactions: Programmable AI meets programmable blockchain
AI agents and robotics are programmable. Smart contacts, the driver of digital assets, are programmable. It makes perfect sense that digital assets would be the preferred payment rail for agent-to-human and agent-to-agent, which includes robotics.
Crypto is an internet-native, programmable money with several advantages for powering the agent-based economy. As AI agents become more autonomous and engage in micro-transactions at scale, crypto’s efficiency, borderless nature and programmability will make it the preferred medium of exchange over traditional fiat rails.
Recent: Sentient open-source AI search outperforms GPT-4o and Perplexity
The true intersection of Web3 and agentic AI for financial transactions could emerge through new tokens and protocols tailored for this use case. These could extend stablecoin capabilities by integrating agent-specific functionalities.
In this scenario, payments could be made using a specialized asset that agents can stake for quality control. Slashing policies could penalize poor performance, while validators could resolve disputes based on task quality.
Additionally, agents’ reputations could be directly tied to their token stakes. Incorporating rules via smart contracts enables users to have control over their autonomous workers/assistants, enabling a shutdown or even a “kill switch,” if necessary, when AI agents start behaving dangerously.
If Goldman Sachs wants to create AI agents that think and act like a seasoned employee in a highly regulated industry and with imperative risk to financial systems and at the extreme financial markets’ stability, it would be vital, not optional, to have these AI agents controlled by programmable tokens.
While this approach requires advancements in both Web3 and agentic AI, it is not as distant as it may seem.
Blockchain development firm Skyfire recently launched a payment platform that allows AI agents to spend money autonomously. Helmed by former Ripple vice president of products and services Amir Sarhangi, the company’s platform enables a business to give a pre-loaded wallet to an AI agent.
The company’s protocol converts the cash into USDC (USDC). In early March, Skyfire brought its payments network that enables AI agents to make autonomous transactions out of beta.
Using digital assets for robotics, VR devices and agentic AI transactions goes beyond a mode of payment for…
cointelegraph.com