The traditional use of fermented and uncooked foods naturally rich in bacteria, the growing interest in alternative food preservation using protective bacteria and the current popularity of new fermented plant products make the consumption of food products containing high levels of live bacteria a common feature of today's diet. Such a practice is perceived as beneficial to health, probably through modulation of the gut microbiota, although the underlying mechanisms are not fully understood. The bacterial density reached in such bacteria-enriched food products is compatible with the production of autoinducers (AIs). These signalling molecules mediate the density-dependent communication system, termed quorum sensing (QS), regulating cooperative behaviors (e.g. biofilm development, enzyme secretion). Therefore, non-peptide AIs from bacteria-enriched foods could modulate the gut microbiota of the consumers. To support this hypothesis, we tested the ability of different bacterial strains used in food microbiology to produce AI-2 (a mixture of borated and non-borated inter-convertible compounds derived from 4,5-dihydroxy-2,3-pentanedione), investigated the presence of AI-2 in some fermented foods and quantified the survival of AI-2 within a modelled food during an in vitro digestibility test. About a hundred bacterial strains with potential as starters, probiotics, or protective bacteria were screened for their production of AI-2, using Vibrio campbellii as a bioluminescent reporter strain. Then concentrations of AI-2 were quantified by mass spectrometry after culture of the most active strains on food matrices derived from milk or fish, as well as in various commercial food products. Finally, we quantified the degradation of AI-2 under conditions mimicking the different phases of digestion in vitro by subjecting cultures of L lactis CIRM-BIA242 conducted on milk ultrafiltrate to the protocol validated in the framework of COST INFOGEST 2.0.Numerous strains, except those affiliated to Lactococcus raffinolactis, Enterococcus gilvus, and Lactobacillus delbrueckii, were able to induce luminescence with maximum intensity observed for some strains of Lactococcus piscium, Lactococcus lactis and Propionibacterium freudenreichii. These latter strains also produced AI-2 when grown on milk ultrafiltrate. (i.e. 82.1 ± 1.0 μM, for L. lactis CIRM-BIA242, 18.3 ± 2.0 for L. piscium 07/R/404, 11.1 ± 2,6 and 4.1 ± 0.5 for P. freudenreichii CIRM-BIA1 and CIRM-BIA121). In line with this, we quantified the highest concentrations (> 40 nmol.g-1) of AI-2 in curd and cottage cheese even though other fermented foods also contained AI-2 in non-negligible concentrations (e.g. > 10 nmol.g-1 for kefir, yoghurt and kimchi). Finally, the residual amounts of AI-2 from milk ultrafiltrate fermented by L lactis CIRM-BIA242 after each phase of the digestion process were 90.0 ± 9.5, 70.6 ± 3.6 and 31.0 ± 9.2% (after oral, gastric and intestinal phases, respectively) of the amounts initiallypresent in the cultures. Despite these partial degradations, the amounts of AI-2 still present after digestion remain significant and consistent with those we measured in the intestinal contents of rats (i.e. 7.0 to 25.5 nmoles.g-1 depending on the digestive segment). Therefore, taken together, our results support a possible role of AI-2 present in bacterial enriched foods in modulating the composition of the gut microbiota of consumers.