Mild cognitive impairment (MCI) describes a period of subclinical cognitive decline associated with an increased risk for the development of Alzheimer's disease (AD). However, research groups have rapidly moved from defining MCI as a risk profile for AD, to promoting MCI as a clinical diagnosis without sufficient evidence of its validity as a diagnostic entity. Specifically, the criterion, subjective change in cognition, remains controversial in light of growing evidence that subjective reports are inadequate predictors of genuine cognitive impairment. Further, operationalisation of the cognitive impairment criterion varies widely across studies in terms of which domains are assessed, and how these domains are assessed. Many studies also fail to acknowledge the psychometric limitations of neuropsychological measures and how this may impact the diagnosis of MCI. In addition, recent research indicates that MCI is longitudinally unstable, with up to 48% of those diagnosed with MCI “recovering” to age appropriate levels of functioning over time. While the MCI concept as it is currently defined may be useful in terms of indicating a risk profile, its ability to sensitively and specifically identify those who will transition to a clinical state remains questionable.