Osteoporosis is a common curable ailment, more commonly seen in geriatric population, particularly prevalent in post-menopausal women. It often presents with generalized bone pain, and fragility fractures, most common in the distal radius, vertebrae, and pelvic bones. When it comes to diagnosing osteoporosis, the best and widely accepted golden standard is the DEXA scan, often interpreted in the form of a T-score and a Z-score of less than −2.5. Though, even in the modern era of evolved imaging, DEXA machines continue to be scarce and are not so profitable to be installed, leading to a massive population of the elderly having untreated and under-diagnosed osteoporosis. CT scan is one of the most common imaging techniques utilized in India for a variety of indications. There has been a lot of buzz regarding the discussions over – “What if one could utilize HU values for diagnosing osteoporosis?” “What if one could predict the risk of fragility fractures without letting one happen?”. A lot of research lately has been focused on employing ML algorithms for achieving this, by using texture analysis of the vertebral body, cortex and trabecular pattern, HU values, and also by using Genant classification of vertebral body fractures. This article aims to review the current situation of opportunistic screening of osteoporosis worldwide and the implications of ML in achieving the same.