AI Found to Underperform Markets Over Long Term, Study Finds

The Facts

A study has found that large language models used for trading purposes do not outperform the market over long periods of time. The research indicates that while AI models may show promising results initially, their performance degrades under changing market conditions. The findings raise questions about the long-term viability of AI-driven trading strategies.

How different outlets are framing this

With only a single source available — the Wall Street Journal — a full cross-outlet framing analysis is not possible. The WSJ's headline and summary frame the story as a limitation or failure of AI trading tools, using language such as 'struggles' to convey underperformance. The emphasis is placed on the long-term horizon and changing conditions as the key variables that expose AI's weaknesses, rather than, for example, highlighting any short-term successes the models demonstrated.

The WSJ, as a financially focused outlet with a sophisticated readership, frames the story in market-performance terms, centering the practical implications for investors and traders. Notably, the framing does not appear to question the study's methodology or offer counterarguments from AI proponents, suggesting a relatively straightforward presentation of the research findings as cautionary for those considering AI-driven investment strategies. Without additional outlets covering the same story, it is not possible to assess whether other publications might have emphasised different angles, such as the broader implications for AI development or regulatory considerations.

Source Articles

• Wall Street Journal25 Jun, 19:00
  AI Struggles to Out-Trade the Market in the Long Term, Study Finds

  Research finds that while large-language models may work well initially, they don’t outperform the market over long periods and in changing conditions