The venture capital (“VC”) industry is fuelled by relationships, networks, and peer comparisons. A function of the limited business and financial track-records of startups, these inputs have unfortunately resulted in a lack of diversity in VC deals and the underfunding of many high potential companies. 

Artificial intelligence (“AI”) is changing the way data is assessed and decisions are made. Increasingly, VC firms, incubators, and accelerators are using AI as a factor in decision-making. Relying on data and objective factors, AI has the potential within the VC industry to shift decision-making towards more informed investment processes that limit unconscious biases. However, AI is only as good as the data it has been given. Since decision-making in the VC industry has historically been influenced by unconscious biases, AI has the potential to produce a similar set of skewed results unless there is a conscious intervention. 

In this problem, there exists an opportunity for VC firms to leverage AI to identify high-potential companies that may otherwise go unnoticed – thus, increasing diversity and diversification within their portfolios and outperforming their peers. To mitigate the risk of biased AI systems, VC firms can develop diverse teams, implement training, and develop algorithms that utilize a multi-stakeholder approach as well as sources of seemingly unconnected data. 

This whitepaper aims to outline the embedded bias challenge associated with AI implementation for VC decision-making and explores best practices for developing AI that can identify high-potential startups that may otherwise be overlooked.