The increasing amount of information online makes it challenging to judge what to believe or discredit. Fact-checking unverified claims shared on platforms, like social media, can play a critical role in correcting misbeliefs. The current study demonstrates how the effect of fact-checking can vary by several factors. We show that fact-checking helps self-correct one's views among young adults. However, this effect is weaker for individuals who perceived the claim negatively at first. Furthermore, borderline messages like “Lack of Evidence” can be perceived as false rather than neutral. We explain these biases via human cognitive mechanisms that avoid risk and uncertainty.