Despite its significance as a contributing factor for late-life dementia risk, illiteracy is frequently underappreciated in the management of dementia. In this study, we estimated the proportion of dementia cases attributable to illiteracy using the population attributable fraction (PAF), and calculated to what extent the monetary cost of dementia could be saved in the future by reducing illiteracy from the South Korean, Latin American, South Asian/Middle Eastern, and African populations. We collected necessary data from the 2011 United Nations Human Development Report and prevalence studies conducted in these regions. Additional variables not included in the above sources were estimated using a logit model under a "trend scenario"-based assumption. Around 16% of the total number of dementia cases in South Korea in 2015 can be attributed to illiteracy, with this figure predicted to decline to around 2% by 2050. This translates to a saving in dementia care costs of approximately 52 billion USD, providing we are successful in theoretically eradicating illiteracy as of 2015, in the population aged 65 years or under. Likewise, reducing illiteracy to 50% in Latin America, South Asia/The Middle East, and Africa by 2050 could generate further cost savings of between 71 and 244 billion, 13 and 94 billion, and 17 and 78 billion USD, respectively. Even public policies aimed solely at reducing illiteracy in the childhood, adolescent, or middle-aged population could potentially have a role in the primary prevention of dementia. Moving forward, governments will need to address this issue in a purposeful and systematic manner.