In a groundbreaking collaboration, Casper Labs and IBM Corp. have joined forces to develop an AI governance tool that harnesses the power of decentralized ledger technology, commonly known as blockchain. This innovative tool aims to introduce version control, monitoring, and access controls for AI models, providing companies with a comprehensive and auditable solution for managing and governing their AI systems.
The need for such a tool emerged as a result of the convergence between AI and blockchain technology, according to Mrinal Manohar, the Chief Executive of Casper Labs. Companies across various industries have recognized the importance of governing AI model behavior and complying with regulatory standards set by government bodies.
Unlike previous attempts to combine AI and blockchain, Casper Labs and IBM are taking a more thoughtful and responsible approach. They acknowledge that AI governance is a critical aspect of modern business operations, not just a superficial addition. Furthermore, they recognize that the rapidly evolving AI industry lacks a well-defined technological framework. This collaboration aims to fill this gap by providing a robust and adaptable solution.
The main feature of this AI governance tool is its utilization of blockchain technology to create a tamper-proof historical record of changes made to AI models. By leveraging distributed ledgers and cryptography, every modification to an AI model version can be documented as a transaction in the blockchain record. This creates a reliable “reset point” that can be referred to in case any issues or discrepancies arise with the model.
According to Manohar, the tamper-resistant and time-synchronized nature of blockchain technology enables highly automated and certifiable AI governance. With the integrated capabilities of IBM’s watsonx.governance toolkit, companies can automate and manage their AI models while adhering to regulatory requirements and addressing ethical and legal concerns.
Furthermore, the combination of blockchain technology and AI governance allows for enhanced version control and transparency. Any deviations or anomalies in AI models can be easily traced back to specific events and changes. This is particularly crucial for addressing issues like hallucinations, where AI models start generating false information more frequently. By analyzing the blockchain record, developers can pinpoint the moment before the problem occurred and revert back to a previous state.
Moreover, large organizations can implement access controls and permissions through blockchain technology, ensuring only authorized individuals or groups can edit or train AI models. This feature is especially valuable for organizations handling sensitive datasets.
Casper Labs and IBM anticipate their AI governance tool to attract customers from highly regulated industries, such as finance, insurance, and supply chain management. These sectors, which rely on accurate and accountable AI systems, will greatly benefit from the transparency and risk mitigation provided by the tool.
The AI governance tool is currently in active beta mode, and Casper Labs is actively seeking feedback from enterprise users to improve its functionality. The official launch of the tool is scheduled for the third quarter of this year.
The source of the article is from the blog elperiodicodearanjuez.es