ABSTRACT Curing HIV-1 infection will require elimination of persistent cellular reservoirs that harbor latent virus in the face of combination antiretroviral therapy (cART). Proposed immunotherapeutic strategies to cure HIV-1 infection include enhancing lysis of these infected cells by cytotoxic T lymphocytes (CTL). A major challenge in this strategy is overcoming viral immune escape variants that have evaded host immune control. Here we report that naive CD8 ⁺ T cells from chronic HIV-1-infected participants on long-term cART can be primed by dendritic cells (DC). These DC must be mature, produce high levels of interleukin 12p70 (IL-12p70), be responsive to CD40 ligand (CD40L), and be loaded with inactivated, autologous HIV-1. These DC-primed CD8 ⁺ T cell responders produced high levels of gamma interferon (IFN-γ) in response to a broad range of both conserved and variable regions of Gag and effectively killed CD4 ⁺ T cell targets that were either infected with the autologous latent reservoir-associated virus or loaded with autologous Gag peptides. In contrast, HIV-1-specific memory CD8 ⁺ T cells stimulated with autologous HIV-1-loaded DC produced IFN-γ in response to a narrow range of conserved and variable Gag peptides compared to the primed T cells and most notably, displayed significantly lower cytolytic function. Our findings highlight the need to selectively induce new HIV-1-specific CTL from naive precursors while avoiding activation of existing, dysfunctional memory T cells in potential curative immunotherapeutic strategies for HIV-1 infection. IMPORTANCE Current immunotherapeutic approaches aim to enhance antiviral immunity against the HIV-1 reservoir; however, it has yet to be shown whether T cells from persons on cART can recognize and eliminate virus-infected cells. We show that in persons on cART a personalized medicine approach using their dendritic cells to stimulate their naive T cells induces potent effector CTL in vitro that recognize and eradicate HIV-1-infected CD4 ⁺ T cells. Additionally, we show that the same stimulation of existing memory T cells results in cytokine secretion but limited effector function. Our study demonstrates that the naive T cell repertoire can recognize persistent HIV-1 during cART and supports immunotherapy strategies for an HIV-1 cure that targets naive T cells, rather than existing, dysfunctional, memory T cells.