How Blockchain and Artificial Intelligence Can Build Trust Together

The issue of trust continues to dominate questions and concerns surrounding emerging technologies such as artificial intelligence and blockchain. Both have already proven to be immensely disruptive to business models, governance, the future of work and other aspects of society. They’ve also both been met with a mixture of enthusiasm and scrutiny, leaving many organizations torn about whether and how to tap into the innovations while avoiding fallout from the risks.

A fortunate plot twist in this story is that when paired together, AI and blockchain stand to reciprocally reinforce trust and reliability. As these technologies are implemented and embedded in more activities and systems, businesses must ensure the tools can be trusted to reliably, consistently and accurately execute the tasks for which they were designed. Human oversight, deliberate strategy and thoughtful technology integrations are required to instill trust in the automation and output produced by AI, uphold the provenance of data that powers advanced systems and manage quality control over time. Blockchain is a natural complementary technology to support these needs.

How does this work? For starters, AI and blockchain are both data-centric systems, so the immutability provided by blockchain can help ensure the quality and integrity of the data feeding into AI systems. Likewise, the machine learning power of AI can ingest and quickly make sense of large pools of data to deliver insights about activities taking place on a blockchain, or accessed via a blockchain and improve data lineage. With that dynamic as the foundation, there are numerous ways these technologies help to reciprocate trust. These include:

Related:Is Farmer Reluctance to Embrace AI Holding Back the Future of Agriculture?

Related:Generative AI Gets to the Root of the MTTR Problem

Early Use Cases

Even in this early state of generative AI, it’s clear that the eventual use cases are likely to span across industries and business functions, touching everything from legal workflows and customer service to marketing, sales automation, research, supply chain management, transactions and more. No matter the use case, though, successful execution will largely rely on data reliability, transparency and trust. With the guidance of technical experts who understand the technical nuances and limitations of these advanced technologies, blockchain and AI can provide that foundation. Example use cases include:

Upholding compliance in cryptocurrency and digital assets offerings 

With the advancement of generative AI to understand text, code and intent, it’s now possible to build tools that can help quickly and accurately answer questions about emerging regulations. These solutions can help organizations efficiently track the changing regulatory landscape across geographies (e.g., MICA in Europe, EU, VARA in Dubai) and answer specific questions about the requirements. For example, tools are in development with FTI Technology’s Blockchain & Digital Assets practice to apply knowledge management techniques such as retrieval augmented generation to query generic and specific questions about the regulatory landscape, cross-referenced with a base of technical and industry knowledge.

This capability can aid digital asset providers and other parties engaging in the ecosystem with compliance and operational readiness in a changing environment. It can also be further extended to the assets themselves. Using knowledge of a proposed transaction, tokens can use these provable AI systems to ensure that the transaction itself will be compliant and only allow a transfer if so.

Blockchain combines knowledge through a shared, immutable ledger from which systems can leverage data from numerous sources. To further ensure outputs or responses are accurate, blockchain proofs or tracking of initial questions and responses can identify the correct AI agent that was used to answer requests. These capabilities can help to create understanding from disparate data and help ensure the correct tools are being used across the environment. Blockchain is a reliable solution to create these cryptographic proofs. They can also be extended into other domains where trust and proof of correctness are required when using AI for information retrieval and summarization.

Transaction monitoring, analytics and understanding

AI has already been established as an effective tool for anomaly detection and monitoring baseline transaction behavior to allow compliance teams to detect behavioral deviations that may indicate fraud. These models can also be applied to analyzing the web of digital wallet interactions for potentially suspicious accounts or laundering pathways. In parallel, blockchain systems support functionality across behavioral datasets to train these AI-powered fraud detection models. All the data that feeds into these models can be stored on a blockchain to maintain an immutable, transparent record that can be used as documentation and a source of evidence during investigations or regulatory inquiries relating to transactions. Records can be further supported with retrieval augmented generation systems to provide context and rate severity of activity, providing compliance teams enhanced insight about potential issues, so they can be investigated and/or resolved quickly.

Data and knowledge access

A key strength of blockchain technology is its ability to integrate numerous systems within one platform. For example, in supply chains, blockchain can be used to bridge enterprise resource planning systems. Or in health care, to permission data access on off-chain platforms. And in finance to automate transactions or information sharing across systems. When integrated with In combination with generative AI and retrieval systems, blockchain can be used to provide permissions and service policies across multiple datasets, thereby creating powerful knowledge platforms, where the information a user can access may be spread across organizations and data providers.

The technology can also serve as the integration layer for multiple data providers, allowing separate parties to apply generative AI queries against specific information as defined by smart contracts. Such a system could allow users to access the relevant data to which they’re entitled and get provable and informed responses accordingly. In transactions or commerce between parties, ownership of data can also be identified, ensuring the right entity is compensated or credited appropriately for their data and that the data is coming from a known, provable space.

As generative AI is used more widely to allow for conversational interrogations of data for both consumers and enterprises, trust remains an issue. Blockchain can be used to verify that the responses are correct and trustworthy. Indeed, these systems can be built to be trusted, but they must be designed with appropriate technical controls, governance fundamentals and strict attention to data quality. Strategy and models must be aligned with organizational principles. Data sets and outputs must be clean, accurate and proportional.

By responsibly leveraging these advancements in tandem with human expertise and oversight, mistakes and biases can be more reliably weeded out and governance can be maintained to ensure the technology does not deviate from intended purposes and outcomes.