The European Commission asked EFSA for a scientific opinion on the risks for animal and human health related to the presence of quinolizidine alkaloids (QAs) in feed and food. This risk assessment is limited to QAs occurring in Lupinus species/varieties relevant for animal and human consumption in Europe (i.e. Lupinus albus L., Lupinus angustifolius L., Lupinus luteus L. and Lupinus mutabilis Sweet). Information on the toxicity of QAs in animals and humans is limited. Following acute exposure to sparteine (reference compound), anticholinergic effects and changes in cardiac electric conductivity are considered to be critical for human hazard characterisation. The CONTAM Panel used a margin of exposure (MOE) approach identifying a lowest single oral effective dose of 0.16 mg sparteine/kg body weight as reference point to characterise the risk following acute exposure. No reference point could be identified to characterise the risk of chronic exposure. Because of similar modes of action for QAs, the CONTAM Panel used a group approach assuming dose additivity. For food, the highest mean concentration of Total QAs (TotQAs) (i.e. the 6 most abundant QAs) was found in lupin seed samples classified as ‘Lupins (dry) and similar‐’. Due to the limited data on occurrence and consumption, dietary exposure was calculated for some specific scenarios and no full human health risk characterisation was possible. The calculated margin of exposures (MOEs) may indicate a risk for some consumers. For example, when lupin seeds are consumed without a debittering step, or as debittered lupin seeds high in QA content and when ‘lupin‐based meat imitates’ are consumed. For horses, companion and farm animals, other than salmonids, the available database on adverse effects was too limited to identify no‐observed‐adverse‐effect levels and/or lowest‐observed‐adverse‐effect levels and no risk characterisation was possible. For salmonids, the CONTAM Panel considers the risk for adverse effects to be low.