Large language models (LLMs) and their integration into agentic and embodied systems are reshaping artificial intelligence (AI), enabling powerful cross-domain generation and reasoning while introducing new risks. Key concerns include hallucination and misinformation, embedded and amplified biases, privacy leakage, and susceptibility to adversarial manipulation. Ensuring trustworthy and responsible generative AI requires technical reliability, transparency, accountability, and attention to societal impact. The present study conducts a review of peer-reviewed literature on the ethical dimensions of LLMs and LLM-based agents across technical, biomedical, and societal domains. It maps the landscape of risks, distills mitigation strategies (e.g., robust evaluation and red-teaming, alignment and guardrailing, privacy-preserving data practices, bias measurement and reduction, and safety-aware deployment), and examines governance frameworks and operational practices relevant to real-world use. By organizing findings through interdisciplinary lenses and bioethical principles, the review identifies persistent gaps, such as limited context-aware evaluation, uneven reporting standards, and weak post-deployment monitoring, that impede accountability and fairness. The synthesis supports practitioners and policymakers in designing safer, more equitable, and auditable LLM systems, and outlines priorities for future research and governance.

This paper introduces Harmonizer, a universal framework designed for tokenizing heterogeneous input signals, including text, audio, and video, to enable seamless integration into multimodal large language models (LLMs). Harmonizer employs a unified approach to convert diverse, non-linguistic signals into discrete tokens via its FusionQuantizer architecture, built on FluxFormer, to efficiently capture essential signal features while minimizing complexity. We enhance features through STFT-based spectral decomposition, Hilbert transform analytic signal extraction, and SCLAHE spectrogram contrast optimization, and train using a composite loss function to produce reliable embeddings and construct a robust vector vocabulary. Experimental validation on music datasets such as E-GMD v1.0.0, Maestro v3.0.0, and GTZAN demonstrates high fidelity across 288 s of vocal signals (MSE = 0.0037, CC = 0.9282, Cosine Sim. = 0.9278, DTW = 12.12, MFCC Sim. = 0.9997, Spectral Conv. = 0.2485). Preliminary tests on text reconstruction and UCF-101 video clips further confirm Harmonizer’s applicability across discrete and spatiotemporal modalities. Rooted in the universality of wave phenomena and Fourier theory, Harmonizer offers a physics-inspired, modality-agnostic fusion mechanism via wave superposition and interference principles. In summary, Harmonizer integrates natural language processing and signal processing into a coherent tokenization paradigm for efficient, interpretable multimodal learning.