Key facts
- A third of investors who buy data report its quality has deteriorated over the past couple of years.
- Some investors blame artificial intelligence for the decline in data quality.
- AI is affecting data quality by being used to generate or map data, or by justifying staff reductions in data quality roles.
- Hedge funds are concerned about complying with regulations when data is generated by AI.
- Some data buyers will use a vendor's AI model if it's free, but not if they have to pay.
Hedge funds are increasingly encountering issues with the quality of data they purchase, with a significant portion of investors reporting a decline over the last couple of years. This deterioration is partly attributed to the growing use of artificial intelligence by data providers. AI is impacting data quality in two primary ways: providers are using it to generate or map the data they sell, or they are using it as a rationale to cut back on human staff dedicated to ensuring data quality. This trend is concerning for hedge funds, which rely heavily on alternative data for their investment strategies. The use of AI in data generation raises concerns about transparency, as funds need to understand how data is created to ensure compliance with federal regulations regarding information availability and privacy. Furthermore, there's a risk that the AI models used by vendors may not be sufficiently robust or accurate. "The data might be there, but the LLM might be rubbish," said Daryl Smith, head of research at Neudata. Hedge funds, often equipped with substantial tech budgets and skilled data science teams, generally prefer to process data through their own systems rather than relying on third-party AI models. The "secret sauce" of many investment strategies involves proprietary data processing techniques. "Funds trust AI with their workflow much more than they trust it with their alpha," Smith added. Daniel Entrup, cofounder of AggKnowledge, noted an increase in basic data hygiene issues as industry focus shifts to AI use cases. His firm, which performs data cleaning services, has seen a significant rise in such work. Buyers express frustration that these AI initiatives are not driven by client demand. "Show me the customer research where clients were asking for it," Entrup stated. Vendors employing more AI in their processes have not seen a corresponding increase in business, with some buyers indicating they would only use AI-driven tools if offered for free. "And it only takes one hallucination to ruin a dataset," said one hedge fund manager.