The black-hole information paradox is usually formulated as a conflict between two semiclassical statements: Hawking evaporation appears to radiate away mass in a nearly thermal manner, while unitary quantum dynamics forbids the destruction of information. In this paper we construct a macroscopic black-hole surrogate inside the discrete, background-free Quantum Graph Emergent Field Theory (QGEFT) and track both horizon-entanglement proxies and the topological fate of explicitly injected information. Starting from a thermalized Unified Phase 3 universe with N = 1024 sites, we force a localized geometric collapse by driving one node to target degree k = 160, then inject a structured quark-like defect whose internal edges are tracked exactly through 100 subsequent Monte Carlo sweeps. The key observable is a horizon cut-size proxy: the number of edges crossing between the black-hole region and the exterior vacuum, together with a normalized entanglement proxy built from that cut. In the present benchmark the horizon cut-size rises from 1027 to a peak of 1030 edges at sweep 1, while the normalized cut-based entanglement proxy rises from 2.7483 to 2.7563, then trends downward with fluctuations to 2.5941 by sweep 100 as the black-hole core shrinks from degree 161 to 132. At the same time, direct tracking of the infalling quark reveals incomplete local reconstruction: only 25% of the tracked edges appear outside the horizon as distinct edges at any given sampled frame, and the best structural-overlap score reaches only 0.3946. The narrow conclusion supported by the current data is therefore mixed. The black hole exhibits a real topological evaporation channel and a measurable Page-like turnover in the horizon entanglement proxy, but local observables remain Hawking-leaning and do not by themselves recover the full infalling structure. In the QGEFT interpretation, this is evidence for fast scrambling rather than literal information destruction: the information is not recovered as a small intact motif, but is dispersed into the global topological microstate of the graph.
Yaniv Cohen (Fri,) studied this question.