We assessed knowledge and practices related to antibiotic use in Kibera, an urban informal settlement in Kenya. Surveys was employed at the beginning (entry) and again at the end (exit) of a 5-month longitudinal study of AMR. Two-hundred households were interviewed at entry, of which 149 were also interviewed at exit. The majority (>65%) of respondents in both surveys could name at least one antibiotic, with amoxicillin and cotrimoxazole jointly accounting for 85% and 77% of antibiotics mentioned during entry and exit, respectively. More than 80% of respondents felt antibiotics should not be shared or discontinued following the alleviation of symptoms. Nevertheless, 66% and 74% of respondents considered antibiotics effective for treating colds and flu in the entry and exit surveys, respectively. There was a high (87%, entry; 70% exit) level of reported antibiotic use (past 12 months) mainly for colds/flu, coughs and fever, with >80% of respondents obtaining antibiotics from health facilities and pharmacies. Less than half of respondents remembered getting information on the correct use of antibiotics, although 100% of those who did reported improved attitudes towards antibiotic use. Clinicians and community pharmacists were highly trusted information sources. Paired household responses (n = 149) generally showed improved knowledge and attitudes by the exit survey although practices were largely unchanged. Weak agreement (κ = -0.003 to 0.22) between survey responses suggest both that unintended learning had not occurred, and that participant responses were not based on established knowledge or behaviors. Targeted public education regarding antibiotics is needed to address this gap.