BlackRock: AI tools like ChatGPT are a ‘revolution’ in investment analysis, able to pull nuance from vast data sets

Artificial intelligence (AI) technologies, such as ChatGPT, are revolutionising the finance industry. BlackRock, a leading asset management company, is optimistic about leveraging these rapidly evolving technologies to assist clients in building superior portfolios. Jeff Shen, co-CIO and co-head of systematic active equity at BlackRock, explains that an AI model using machine learning can process vast amounts of data. This data can help investment analysts evaluate economic conditions. The data comes from various sources, including daily news, broker reports, expert insights, and government statistics. Additionally, geospatial information can be used to identify trends by analysing patterns, such as the movement of trucks in and out of a company's warehouses or foot traffic in a city. Shen believes that the advent of big data and AI is transforming every industry, and the investment sector will not be an exception. AI tools based on large language models (LLMs), such as Microsoft-backed OpenAI’s ChatGPT and Baidu’s Ernie Bot, have gained global attention this year. Businesses are eager to harness the potential of these technologies. Shen leads a team of experts, including computer scientists, engineers, physicists, mathematicians, and traditional finance analysts. They are already using LLM-based AI to help construct portfolios. BlackRock is not merely using off-the-shelf AI but has invested in developing finance-focused language models. These models are trained on text specific to financial markets and investing. Shen emphasises that the input used to train the machine-learning algorithm is as crucial as the large language model itself. Such technology can help assess vast amounts of data and detect nuanced information, which ultimately aids in understanding market conditions, identifying investment themes, and spotting opportunities to improve investment outcomes. According to a recent study, using LLMs in data analysis costs less than 1% as much as hiring a human analyst while delivering comparable performance.