Background
Generative AI is everywhere. The world’s brightest minds are racing to excel in what is expected to be the single most important game changer since the internet. The industry is on the move:
- The past six months have seen a number of new or updated large language models launched by major industry players. There has also been a strong push for text-to-image and multi-modal models.
- A host of new players are emerging in specific application areas such as research, corporate marketing or pharmaceuticals.
- Companies around the world are exploring how and where to integrate generative AI into their existing business models and workflows.
This is driving both demand and development across the entire value chain, from the makers of chips used for cloud computing, scaling and building up of data centre capacity, to the designers of apps that enable low-level access for end users. Even new job profiles are emerging, such as prompt engineers, AI trainers and AI consultants.
Globally, regulators are striving to keep up with the rapid pace of the industry. Numerous regulatory initiatives focused on AI are gaining traction at a global level. As UK Prime Minister Rishi Sunak hosted the AI Safety Summit in November, US President Joe Biden issued an executive order to establish AI safety and security standards. A few weeks later, the EU institutions agreed on the AI Act. But it is not just newly enacted regulations that create a need for compliance. There are also a number of existing sets of rules, such as intellectual property and data protection laws, to which AI models and service providers are already subject today.
M&A Environment
The rapid rise of generative AI will have a significant impact on the M&A environment. Generative AI has the potential to change today's business models across many sectors:
- Value chains could be disrupted as generative AI tools are adopted by traditional industries.
- At the same time, entirely new products, sectors and markets are emerging. This is driving the entry of new players.
The momentum created by generative AI has already triggered a new wave of M&A activity. Beyond traditional M&A, cross-industry collaborations are on the rise.
M&A structuring
In an AI-driven environment, addressing the classic M&A challenges of protecting key value drivers and ringfencing as well as allocating risks will require an adaptation of established strategies.
The founder-owner business structure of many new AI players brings familiar challenges, such as a lack of institutionalised governance, a non-traditional founder-centric organisation and complex financing arrangements. The structural response could be a more flexible allocation of risk and reward, e.g. through partial rather than full acquisitions, alternative pricing structures such as earn-outs and incentive programmes, and reciprocal contingent call/put option structures.
The specific risk profile of AI centric deals or companies using AI will entail additional challenges, which may need to be reflected in potential new warranties, tailored pre-closing undertakings and/or closing conditions:
- Data protection, intellectual property rights as well as regulatory issues will typically play a key role. These issues often do not lend themselves to a purely documents-based due diligence process and require extensive questioning of management teams. In particular, the data used to train and prompt the underlying AI model will come under increased scrutiny during due diligence. Contractual allocation of risks related to the quality, integrity, ownership and usability of the data used will be key to minimising future liability risks.
- Given the legislative efforts in the regulatory area in all major jurisdictions, due diligence will also need to include compliance checks of the AI models against the current and expected AI-specific regulatory environment and stress-testing the maturity of the target’s AI governance processes. This applies in particular to the way in which the AI model has been designed, as well as potential transparency, risk management and documentation requirements. Similarly, the output of the AI model will need to be specifically reviewed for ownership, protectability, potential infringement of third party rights, and regulatory and reputational risks.
- From a commercial perspective, given fierce competition and fast market dynamics, interchangeability, i.e. the ability to move the product to another AI model, may be particularly important to mitigate business continuity risks associated with specific AI models. Talent retention may also play a key role in deal structuring.
For further details please see Generative AI: Five things for lawyers to consider and GenAI: What are the risks associated with its adoption and how do you mitigate those to maximise the opportunities it offers?
M&A process
The M&A process will benefit from generative AI on several levels. Machine learning tools have already been used extensively in recent years, for example to redact personal data and identify certain contractual clauses during due diligence. Generative AI has the potential to take this to a whole new level, e.g. through enhanced contract review and the preparation of entire work products. As with any particular AI solution, the amount and quality of data on which the AI model is trained, as well as the usual contextual adjustments, will play a key role in the effectiveness of the tools. As a result, an excellent knowledge base and market-leading skills in bridging the gap between law and technology will be more important than ever.
Notwithstanding the expected productivity leap driven by generative AI, human advice and interaction will continue to determine the success of M&A processes. This is not only due to the inherent limitations of generative AI models at this stage (e.g., the potential for hallucinations, factual errors, biases), which require expert cross-checking. More importantly, as long as deals are done by humans, professional experience, judgement and negotiation skills will continue to make a critical difference in successful M&A processes.
Wrap-up
Generative AI will transform the M&A landscape. New buyer-target constellations will require an adaptation of risk mitigation and asset protection strategies. More than ever, data, IP, regulatory and compliance risks will need to be at the heart of any due diligence. Generative AI has enormous potential to facilitate the M&A process. Still, human advice and decision-making will continue to be a key success factor for future M&A transactions.