Artificial Intelligence of Things (AIoT) combines Artificial Intelligence (AI) and Internet of Things (IoT) to enable intelligent, real-time decision-making in applications such as healthcare, smart agriculture, and industrial automation. However, ensuring data security in AIoT environments is challenging due to the limited computational resources, memory, and energy constraints of edge devices. Traditional cryptographic algorithms are often too heavy for such settings, while existing lightweight schemes may lack adequate security. This paper proposes LCE-AIoT, a lightweight cryptographic framework based on an optimized version of the Ascon algorithm. Designed for 128-bit authenticated encryption, LCE-AIoT reduces computational overhead by minimizing permutation rounds and memory usage, making it suitable for microcontrollers like ESP32 and STM32. The framework supports both pre-shared and ECC-based key exchange and integrates seamlessly with lightweight IoT operating systems such as Contiki-NG and TinyOS. Experimental results show that LCE-AIoT achieves superior performance compared to existing lightweight algorithms including PRESENT, SIMON, SPECK, and standard Ascon, offering lower execution time, reduced memory footprint, better energy efficiency, and higher throughput. This makes it ideal for secure, low-power AIoT applications. Future enhancements include post-quantum resilience, hardware acceleration, and blockchain integration, ensuring the framework’s adaptability to emerging intelligent systems.