This paper is in press at Journal of Experimental Psychology: Applied (01/10/21). Preregistration: [Study 1][1], [Study 2][3], [Study 3][2], [Study 4a][6], [Study 4b][7], [Study 5][8], [Supplementary Study 1][4], [Supplementary Study 2][5] Abstract: As technology advances, people increasingly outsource cognitive tasks and can more easily access others’ knowledge. While externalized aids often support human abilities, they may also make it more difficult for people to assess their own competence. Indeed, using online search engines leads people to treat searchable information as if they already know it (Fisher, Goddu, & Keil, 2015). Six primary and two supplemental studies (N = 3,262) extend previous research by exploring how illusions of knowledge result from reliance on other agents. After teaming with knowledgeable partners (artificially intelligent agents or human teammates) on a trivia quiz, people overestimated how well they would perform on future quizzes for which help was not available; this bias was not evident for participants who never received help. Moreover, overconfidence was insensitive to whether assistance was provided on hard vs. easy problems, or even whether the assistance was accurate. Receiving outside assistance creates an ambiguity regarding who deserves credit for success or blame for failure. When this ambiguity is removed, people become better calibrated. These results indicate that reliance on technology and outside knowledge may change our view of ourselves— convincing us we are more capable than we really are. [1]: https://osf.io/vfa5x/?view_only=e39c51ff7d104c2dadc90561bb920d5b [2]: https://osf.io/ftvxu/?view_only=1933114068c348e399f898914d05c5df [3]: https://osf.io/ebm9w/register/5771ca429ad5a1020de2872e [4]: https://osf.io/ygbtj/register/5771ca429ad5a1020de2872e [5]: https://osf.io/7b8vt/register/5771ca429ad5a1020de2872e [6]: https://osf.io/rqc5e [7]: https://osf.io/46s5t [8]: https://osf.io/he37v