Artificial Intelligence‐Integrated Microbial Fuel Cell: A Self‐Sustained Biosensor for Predictive Heavy Metal Toxicity Detection

ABSTRACT This study reports the development of a self‐sustained single‐chamber microbial fuel cell (SCMFC) biosensor integrated with artificial intelligence (AI) for rapid assessment of heavy metal toxicity in water. The anode was modified with a conductive and biocompatible polyindole/titanium oxide/molybdenum oxide (PIn/TiO 2 /MoO 3 ) nanocomposite to enhance electron transfer and microbial attachment. The biosensor was evaluated against four common toxic metals Hg 2+ , Cr 6+ , Cu 2+ , and Ni 2+ at concentrations ranging from 2 to 20 mg/L. Concentration‐dependent inhibition of microbial activity was observed, with mercury exhibiting the highest toxicity (88.88% inhibition at 20 mg/L), followed by chromium (81.81%), copper (71.42%), and nickel (27.77%). A hybrid Artificial Neural Network–Genetic Algorithm (ANN–GA) model was employed to predict toxicity based on inhibition ratios, achieving superior accuracy ( R 2 = 0.998; RMSE = 0.262 mg/L) compared to baseline ANN models ( R 2 = 0.974). The biosensor demonstrated excellent reusability through regeneration using sodium acetate (1 g/L), with complete voltage recovery (0.01–0.35 V) achieved within 70–110 h depending on metal type. These results establish the PIn/TiO 2 /MoO 3 @Toray carbon modified SCMFC as a robust, cost‐effective, and AI‐enhanced biosensing platform capable of real‐time, quantitative monitoring of heavy metal toxicity in wastewater.