Given the continuing heavy toll of the COVID-19 pandemic and the emergence of the Delta (B.1.617.2) and Omicron (B.1.1.529) variants, the WHO declared both as variants of concern (VOC). There are valid concerns that the latest Omicron variant might have increased infectivity and pathogenicity. In addition, the sheer number of S protein mutations in the Omicron variant raise concerns of potential immune evasion and resistance to therapeutics such as monoclonal antibodies. However, structural insights that underpin the potential increased pathogenicity are unknown. Here we adopted an artificial intelligence (AI)-based approach to predict the structural changes induced by mutations of the Delta and Omicron variants in the spike (S) protein using Alphafold. This was followed by docking the human angiotensin-converting enzyme 2 (ACE2) with the predicted S proteins for Wuhan-Hu-1, Delta, and Omicron variants. Our in-silico structural analysis indicates that S protein for Omicron variant has a higher binding affinity to ACE-2 receptor, compared to Wuhan-Hu-1 and Delta variants. In addition, the recognition sites of the receptor binding domains for Delta and Omicron variants showed lower electronegativity compared to Wuhan-Hu-1. Importantly, further molecular insights revealed significant changes induced at fusion protein (FP) site, which may mediate enhanced viral entry. These results represent the first computational analysis of structural changes associated with Omicron variant using Alphafold, Collectively, our results highlight potential structural basis for enhanced pathogenicity of the Omicron variant, however further validation using X-ray crystallography and cryo-EM are warranted.