Businesses across industries are transforming into data enterprises and investing at record rates in AI-focused acquisitions. Yet, neither the risks nor the opportunities presented by these acquisitions are adequately assessed by traditional due diligence.
Risks such as algorithmic discrimination, lack of transparency and unreliable performance are increasingly the causes of AI failures. Critically, the start-up’s culture — the values, people and processes that govern its use of data and AI — is arguably the best indicator of its long-term value as an acquisition. Clarity on these criteria is needed for both acquirers and startups, from the beginning of their conversations.
Therefore, the Data & Trust Alliance has created Responsible Data & AI Diligence for M&A, a new tool for use by M&A teams in their target screening and due diligence to assess the value and risks of data, algorithms and the cultures in which they are built.
Suggested for the target-screening process, it helps an acquirer assess a target's mindset around data and AI and the mechanisms in place to sustain a culture of responsibility and rigor. Areas of inquiry include business purpose; values in practice; and processes to detect, mitigate and monitor data and AI issues. The module explores how the target’s teams work—for example, whether a learning mindset is incentivized and how trade-offs are made.
Assesses how data is sourced, used and responsibly governed, in order to understand its true value and utility for an acquirer and whether any mitigation is required. It inquires into data quality, data bias, data consent and rights, including third-party usage rights.
Assesses the design, deployment and monitoring of algorithmic models to ensure they perform as intended and minimize unintended consequences. It includes inquiries into a target’s approach to sourcing and managing training data, explainability, robustness, fairness, performance monitoring, and independent audits.
Data Diligence and Algorithmic Diligence supplement an organization’s existing technology, privacy and security diligence.
More than 80 experts contributed to the development of the tool. First, a cross-Alliance team of member company experts and external specialists in AI ethics and policy, AI risk, legal and compliance, data quality and diligence, and mergers and acquisitions came together to create the new criteria and associated education and guidance. The work was then tested and refined with input from additional Alliance member company experts and external leaders in corporate development, data, AI and technology ethics.
As of October 20, 2022, 19 members of the Alliance are adopting all three Responsible Data & AI Diligence for M&A modules and adapting for use within their current due diligence.
Download the Responsible Data & AI Diligence Overview here. Please contact us to review the diligence modules and learn more.
To guide development of the Responsible Data & AI Diligence tool for M&A, the Alliance collaborated with distinguished experts in law relating to corporate mergers and acquisitions and in AI risk.
Cravath has been a premier U.S. law firm for two centuries. Each of its practices is highly regarded, including its preeminent M&A and capital markets practices and broad‑based intellectual property and digital assets practice.
Founded in 2020, Credo AI is a venture-backed company that brings context-driven governance and risk assessment to ensure responsible – compliant, fair, transparent and auditable – development and use of AI.