Vol. 7 No. 06 (2025)

The exponential growth of digital content has driven the need for more intelligent, context-aware information retrieval systems. While traditional keyword-based search engines remain foundational, they often fall short of capturing deeper semantic meaning. This article explores the evolution, methodologies, and recent developments in intelligent information retrieval systems powered by artificial intelligence. Special attention is given to the use of machine learning, natural language processing (NLP), and neural networks to improve relevance, personalization, and contextual understanding, including the application of learning-to-rank techniques. The paper contrasts the strengths and limitations of conventional search technologies with those of AI-driven models. A critical part of the study focuses on potential risks associated with AI-based search engines, including environmental concerns linked to the heavy water consumption of data centers relying on water-based cooling systems. The research concludes that a holistic approach is needed in the design and implementation of AI-powered search systems—one that integrates ethical, cognitive, and environmental considerations. This article will be of interest to professionals in media and information technology, researchers, and developers engaged in building intelligent search infrastructures.

KRAS mutations play a key role in the pathogenesis of acute myeloid leukemia (AML), occurring in 10–15% of cases and being associated with aggressive disease progression and therapeutic resistance. Despite significant advances in the treatment of KRAS-mutant solid tumors, including the approval of allele-specific G12C inhibitors, the potential of pan-KRAS inhibitors in hematologic malignancies remains insufficiently explored. This study evaluates a pan-KRAS inhibitor structurally analogous to BI-2493 in the SKM-1 cell line model (KRAS G12D+). In vitro results demonstrate reduced cell viability, induction of apoptosis (Annexin V+), and suppression of the KRAS–MEK–ERK signaling cascade. The findings are contextualized with data from Popow et al. and Revvity/Boehringer Ingelheim, enabling a comparative analysis of G12D-mutant model sensitivity across tumor types. The discussion addresses the potential for in vivo xenograft testing, combination strategies with SHP2 and BCL2 inhibitors, and the application of PROTAC degraders as alternative approaches in resistant settings. These results provide the first evidence of pan-KRAS inhibitor efficacy in an AML model, highlighting its relevance for targeted therapy in hematologic malignancies and supporting further preclinical investigation aimed at integration into personalized oncology protocols.