JP Morgan, citing the IMF, warned AI is increasingly used for risk pricing,
credit allocation and trading decisions. While this can improve execution
efficiency and liquidity in normal conditions, widespread reliance on similar
data, models and signals can produce highly synchronized de-risking in stress —
simultaneous selling and credit tightening depress prices and liquidity, which
then reinforces model sell signals in a procyclical feedback loop. The IMF says
the systemic threat comes less from isolated model errors than from many models
making similar, seemingly rational decisions, increasing market correlation and
the risk of amplified volatility and liquidity shortfalls.