Over the past three decades, the definition and diagnostic boundaries of Alzheimer’s disease (AD) have been repeatedly revised due to significant progress in understanding of the pathogenesis of neurodegeneration associated with Alzheimer’s disease and development of high-tech diagnostic methods. The current approach to diagnosing AD is based on the discovery of biomarkers that reflect the two main neuropathological processes involved in the development of primary neurodegeneration underlying AD-abnormal amyloidogenesis and neuronal degeneration. The currently available diagnostic tools are limited to the detection of cerebrospinal biomarkers and/or assessment of the abnormal amyloid and tau protein burden in the brain via amyloid and tau positron emission tomography ligands. Practical implementation (mostly in the research field) of the biological model of AD diagnosis has led to a significant expansion of its diagnostic boundaries with the inclusion of predementia AD stages: asymptomatic and symptomatic, the latter is clinically corresponding to amnestic mild cognitive impairment (aMCI-amnestic mild cognitive impairment). On the one hand, this approach significantly expands the possibilities to study and use preventive technologies aiming to avert or delay the progression of predementia cognitive impairment to dementia but, on the other, it is associated with a number of negative implications from both the clinical and ethical points of view. A significant limitation of purely biological diagnosis of AD based on biomarker levels is due to the low prognostic value of biomarkers, which can cause diagnostic confusion in certain circumstances. Moreover, since the future evolution of the asymptomatic stage is not yet clear and there are still no reliable ways to prevent the cognitive and behavioral symptoms associated with AD, disclosure of stressful information about this “terrifying” diagnosis to patients can cause irreversible damage by triggering depressive disorder, which is a risk factor of AD itself. The current knowledge about AD prognosis in amyloid-positive cognitively unimpaired patients is insufficient. The most adequate approach to early AD diagnostics appears to be the clinical and biological model, as recommended by the International Working Group (IWG2021), which requires a combination of the clinical AD phenotype and the detection of biomarkers specific to this disease. The article discusses the potential directions for the development of biological diagnostic methods, including those based on the so-called peripheral (plasma) biomarker technologies and promising directions for the development of biological methods of secondary AD prevention.