Direct detection of highly reactive, atmospheric hydroxyl radicals (OH) is widely accomplished by laser-induced fluorescence (LIF) instruments. The technique is also suitable for the indirect measurement of HO2 and RO2 peroxy radicals by chemical conversion to OH. It requires sampling of ambient air into a low-pressure cell, where OH fluorescence is detected after excitation by 308 nm laser radiation. Although the residence time of air inside the fluorescence cell is typically only on the order of milliseconds, there is potential that additional OH is internally produced, which would artificially increase the measured OH concentration. Here, we present experimental studies investigating potential interferences in the detection of OH and peroxy radicals for the LIF instruments of Forschungszentrum Jülich for nighttime conditions. For laboratory experiments, the inlet of the instrument was over flowed by excess synthetic air containing one or more reactants. In order to distinguish between OH produced by reactions upstream of the inlet and artificial signals produced inside the instrument, a chemical titration for OH was applied. Additional experiments were performed in the simulation chamber SAPHIR where simultaneous measurements by an open-path differential optical absorption spectrometer (DOAS) served as reference for OH to quantify potential artifacts in the LIF instrument. Experiments included the investigation of potential interferences related to the nitrate radical (NO3, N2O5), related to the ozonolysis of alkenes (ethene, propene, 1-butene, 2,3-dimethyl-2-butene, α-pinene, limonene, isoprene), and the laser photolysis of acetone. Experiments studying the laser photolysis of acetone yield OH signals in the fluorescence cell, which are equivalent to 0.05 × 106 cm−3 OH for a mixing ratio of 5 ppbv acetone. Under most atmospheric conditions, this interference is negligible. No significant interferences were found for atmospheric concentrations of reactants during ozonolysis experiments. Only for propene, α-pinene, limonene, and isoprene at reactant concentrations, which are orders of magnitude higher than in the atmosphere, could artificial OH be detected. The value of the interference depends on the turnover rate of the ozonolysis reaction. For example, an apparent OH concentration of approximately 1 × 106 cm−3 is observed when 5.8 ppbv limonene reacts with 600 ppbv ozone. Experiments with the nitrate radical NO3 reveal a small interference signal in the OH, HO2, and RO2 detection. Dependencies on experimental parameters point to artificial OH formation by surface reactions at the chamber walls or in molecular clusters in the gas expansion. The signal scales with the presence of NO3 giving equivalent radical concentrations of 1.1 × 105 cm−3 OH, 1 × 107 cm−3 HO2, and 1.7 × 107 cm−3 RO2 per 10 pptv NO3.