Understanding the credibility of AI in the context of an investment is an increasingly common challenge. Particularly as AI encompasses such a wide array of technologies and deployments.
We recently advised on a transaction where the investment thesis was founded on an AI capability presented in the Information Memorandum. We realised that the data set being used was unable to support the predictions the AI was supposed to have. This had a material effect on the valuation and our client was able to use this information to reevaluate the deal.
If you’re considering an AI-enabled investment some good questions to consider when assessing the asset are:
- How is training data used and what’s the risk of bias?
- To what degree are the roadmap and technology documented?
- Are the tools appropriate for the chosen techniques?
- What is the level of performance, how is it measured and how might it scale?
- Are there any regulatory and ethical considerations?