The detection of gravitationally induced entanglement does not serve as theory-independent proof of the nonclassicality of the gravitational field. This is the conclusion of a new paper in Physical Review D, published on 19 October 2026. It was previously believed that the very fact of entanglement between two massive objects via gravity automatically points to the quantum nature of the field, but that certainty has now been undermined.
The authors of the paper examine the so-called "no-go" theorems that supposedly forbid classical gravity from creating entanglement. It turns out that these theorems rest on the assumption of locality of interactions, which in the context of gravity looks artificial. Even quantum electrodynamics and perturbative quantum gravity violate this locality in the same sense. Thus, a classical gravitational field may well induce entanglement if one abandons the strict requirement of locality.
Imagine two massive bodies in a superposition of positions: their gravitational interaction creates a relative phase that can lead to entanglement. It was previously thought that only a quantum field is capable of this without violating locality. It is now clear that a phase shift can also arise in a semiclassical picture, where the field remains classical but the interaction is nonlocal by information criteria.
This changes the entire strategy of tests of quantum gravity. Instead of searching for a "smoking gun" in the form of entanglement, one will have to carry out a quantitative comparison of the predictions of different models in the regime of low-energy quantum gravity. It is necessary to measure not merely the presence of an effect, but its magnitude, its dependence on parameters, and possible deviations from classical expectations.
The paper emphasizes that the search for other signatures — non-Gaussianity, graviton noise, detection of individual gravitons, or thermodynamic effects — becomes even more relevant. Only the combination of such observations will make it possible to distinguish a quantum gravitational field from a classical one.
The results, published in Physical Review D on 19 October 2026, show that experimenters will have to develop more subtle protocols and compare data with specific theoretical calculations, rather than relying on a single qualitative feature.
