Points to note in M&A deals involving AI businesses
The pace of innovation and transformation driven by Artificial Intelligence (AI) has made some traditional considerations in mergers and acquisitions redundant. Deals involving AI businesses could give rise to specific risks. Buyers may want certain protections to mitigate against these. Despite the clear benefits of AI, there might be questions around its reliability, performance, and long-term sustainability which heighten the need for carefully crafted warranties to protect investments and to ensure the delivery of promised functionalities.
As AI businesses increasingly become targets in M&A transactions, it is important that deal advisors are aware of the particular risks involved and the evolving regulation affecting the sector.
Approaching AI deals
Parties should approach AI deals as follows:
- Buyers should conduct thorough due diligence on the AI technology, its capabilities, limitations, and potential liabilities, as well as the sellers' track record with the technology (including any evidence of their prior AI research and development), reputation, and financial stability. This will help to identify any red flags, gaps, or inconsistencies in the warranties and to adjust the deal terms accordingly.
- Sellers should be prepared to provide sufficient evidence and documentation to back up their AI specific warranties, such as test results, audit reports, certifications, licenses, and contracts. They should also be transparent about any known or anticipated issues, challenges, or limitations of the AI technology and how they plan to address them.
Buyers should conduct thorough due diligence on the AI technology, its capabilities, limitations, and potential liabilities, as well as the sellers' track record with the technology.
Although the technology is not new, the regulation of it is. The parties should consider the implications of any changes in the regulatory, legal, or ethical landscape that may affect the AI technology or its use. They should include provisions for updating, modifying, or terminating the warranties in case of such changes, as well as dispute resolution mechanisms and indemnities for any breaches or losses arising from the warranties.
Dealing with risks through warranties
The AI-related risks identified in the due diligence phase should be addressed in the share purchase agreement through appropriate warranties and indemnities, signing, or closing conditions. These due diligence findings may also affect the valuation, negotiation, and structuring of the M&A transaction. Such types of warranties will also need to be considered when approaching AI deals:
- Well defined parameters: To ensure AI systems meet specified performance levels, warranties committing to accuracy, consistency, and reliability within defined parameters should be included. Post-deployment monitoring, bias removal, and, if necessary, human review for riskier outputs should be outlined. Compatibility with existing infrastructure, scalability, and adherence to quality and security standards for algorithms, models, systems, and infrastructure are crucial considerations.
- Intellectual Property (IP): Protection of IP is paramount, and sellers should warrant ownership or valid licenses for all relevant IP rights. Non-infringement of third-party rights, proper and maintained protection of IP assets, and disclosure of any pending or threatened claims related to IP must be assured.
- Technical maintenance: Maintenance and support warranties should include guarantees to carry out consistent software updates, patches, and repairs, along with timely and effective technical support for malfunctions, errors, or user issues. This is essential for the optimal functioning of AI systems.
Sellers should warrant providing insights into how the AI system arrives at decisions and implementing measures to mitigate biases for fairness and non-discrimination.
- Mitigating risks: Risk warranties address the identification and mitigation of potential risks, including impact assessments, updates, steps taken to rectify identified risks, clarity on future risks, and assurances against reputational damage from the technology.
- Data: Data security, privacy, and governance warranties are critical due to AI's significant reliance on valid data. These warranties should cover robust data protection measures, compliance with data protection laws, and ethical considerations in data collection, storage, and processing.
- Decision making: Transparency and explainability warranties address the challenge of understanding AI decision-making processes. Sellers should warrant providing insights into how the AI system arrives at decisions and implementing measures to mitigate biases for fairness and non-discrimination.
- Sustainability: Long-term viability warranties are necessary to address concerns about the sustained functionality of AI technology. This includes commitments to ongoing development and clear exit strategies in the event of product discontinuation and minimising disruptions for users.
