For AI to revolutionise portfolio management, it is important to have not only an abundance of data, but also the ability to create and collect new investment data as and when it is created.
And once it’s collected, the challenge is then to cut through the ‘noise’ that so famously infects market information. So, does this data exist in the abundance needed to enable artificial intelligence to predict stock market movements? And to what extent can machine-learning models be used to carry out complex computational tasks and data analysis?
Speaking at the inaugural FundsTech Forum in London on Thursday, Caroline Minio-Paluello, chief executive of Arabesque AI, said that using AI to power fully investments would be challenging.
“I can definitely tell you it’s not easy. You still need to deliver performance. The whole world needs to adapt to have more adaptability, and it’s far from being the magic solution that everyone has figured out how to make money from.”
Greater automation efficiency in AI is needed, said Minio-Paluello. “This is not a replacement game. It’s an augmentation game. And it’s the question of how you manage that.”
Nabil Cherrat, head of product strategy investment at French funds giant Amundi, said that, despite the complexity involved and challenges around data, accuracy and privacy, AI could be useful in helping to predict market movements and returns with the aim of out-performing the market.
“We need to solve those challenges first and this is what we need to focus on first before going further.”
