Clifford Chance

AI has quickly become one of the biggest stories for Technology investors so far this year, snowballing steadily since the launch of OpenAI's ChatGPT at the end of 2022.

The arrival of 'generative AI' that can produce new content (text-based responses to queries, images, audio, video and even code) marks a new and exciting phase in the development of mainstream AI applications. The starting gun has been fired for a race to capture the market.  

ChatGPT has been a cultural phenomenon. 

ChatGPT set the record for fastest growing user base having reached 100 million monthly active users in just two months after launch. Analysts are seeing AI as an existential threat to incumbent search engines, providing direct text-based responses to queries rather than search results in the form of links. Microsoft was the first search giant to announce that it is integrating OpenAI technologies into its Bing search engine, with Google quickly responding by rolling out "Bard", its experimental AI service.

We expect to see a continued boost to investment activity in the sector.

AI investments have been capital intensive to date and will continue to need  greater access to capital. In part, this is due to the tremendous cost of developing, training and running AI.  AI requires access to cloud servers and data centres. OpenAI started with several billion dollars from its original investors in 2015 and by 2019 Microsoft had invested another $3 billion and have just committed another $10 billion for a third phase partnership. Each ChatGPT query is said to cost a few cents, however when applied across 100 million users with multiple queries per day the operating expenses accumulate rapidly (although subscription and premium models are now being used to generate revenues).

The impact of AI will be felt well beyond the limits of internet search

The technology has been quickly finding use cases across industries that gain an edge by crunching large data sets to derive predictions, from pharmaceuticals to financial services. AI will be a tool that frees up its human operators from the mundane or repetitive tasks to concentrate on higher-value tasks. As an example,  Copilot (developed by GitHub and OpenAI) was launched as an AI assistant that suggests code and functions to software engineers in real time (see our article: Trouble in the cockpit? Popular AI tool faces class action lawsuit). AI will also impact media and creative industries, lowering the cost of producing new content at massive scale but creating a premium for media providers able to cut through the noise with high-quality and curated content. Experts predict that by 2026, as much as 90% of all internet content will be generated by AI.

Some key legal implications of acquiring or investing in AI businesses:

Companies and investors committing to AI need to be fully-aware of the complex and evolving legal environment regulating AI and data, which can impact an investment's risk profile, reputational risk and, by implication, the future value and return of their investment:

• AI needs data. Data is a highly-regulated asset and AI needs large quantities of it, whether in the form of training sets at the development stage or as inputs for the up and running algorithm to process. Investors will need to have a clear AI data strategy to be certain that they are able to use their existing data (e.g. customer data) and continue to source new data from third-parties.  Investors need to consider whether the data has been collected in compliance with law and who bears the risk in relation to third-party arrangements.
• AI specific regulations. Investors need to consider whether the rapidly changing regulatory landscape, for both data and AI, impacts the risk profile of the target business. Investors should be familiar with the regulations that are expected to come into force soon (e.g. the EU's AI Regulation) as well as which regulators have jurisdiction over aspects of the target business. Regulators are already focused on unacceptable risks from AI systems such as where they distort human behaviour, exploit vulnerable groups, relate to 'social scoring' by public authorities,  or involve real-time biometric identification.  Investors should also consider what human oversight has been put in place to monitor and, if necessary, override the AI system. Businesses should ensure that they have AI policies and systems in place now to avoid the accumulation of a legal and operational 'technical debt' that will be expensive to rectify and repair in future.
  
• Reputational risk. Although many of the potential risks associated with AI technologies are addressed and mitigated by incoming AI regulations,  investors should nonetheless consider whether the target has self-regulation in place for ethical questions that are not (yet) covered by law but may nonetheless lead to long-term reputational damage. Industry experts have already noted the tendency of large language models (LLM) to be "confidently incorrect" with a tendency to assert falsehoods as fact. Also, the huge demands for computing power means that AI technologies are likely to be responsible for increased energy use and emissions, raising ESG considerations.
• Cyber security. AI programs are treating huge volumes of potentially high-risk and sensitive data (such as customer information, patient data, or financial data).  As well as considering how this was collected, where it was sourced from and other data protection compliance points, investors need to understand the target's cybersecurity risk profile and who bears the risk of any hack. It is important to consider the information security controls and policies of the target business, whether they are compliant, whether any cyber incidents have occurred previously and how they were handled, and the response plan in place in case of any incident in the future.

Conclusion

The AI-led technological revolution will impact all industries. AI development is a rapidly moving and attractive sub-sector; companies and investors wanting to invest need to be well-advised on the impact for both their existing portfolio and any future strategic investments.