Opinion | 3 spectres haunting global finance must be seen before it’s too late

Large-language models generally rely on similar data, signals and optimisation logic. There’s a collusion effect. Feedback loops heighten the risks of synchronised and repetitive errors at frenzied speeds.

Another concern is position bias, a phenomenon whereby AI overemphasises information at the beginning and end of a text, which Massachusetts Institute of Technology researchers recently uncovered. Data can be invariably conflicting, error-ridden, imprecise, cluttered and ambiguous. Even if gaps are filled, applicability is limited since each event’s causes and complexities are unique.

“AI’s ability to respond swiftly and decisively – combined with its opaque decision-making process, collusion with other engines, and the propensity for hallucination – is at the core of the stability risks arising from it,” Danielsson writes. He warns that “AI crises will be fast and vicious”.