Secure Intellicise Wireless Network: Agentic AI for Coverless Semantic Steganography Communication

TL;DR

This work tackles semantic eavesdropping in SemCom by introducing AgentSemSteCom, a coverless semantic steganography framework powered by agentic AI. It integrates semantic extraction, a digital token to deterministically seed reference image generation, and a diffusion-based coverless scheme with EDICT-based exact inversion and decoupled multimodal guidance, all without relying on private semantic keys or original cover images. Evaluations on UniStega show improved pixel-level fidelity (higher PSNR, SSIM) and lower distortion (MSE, LPIPS) relative to SemSteDiff, while token-based perturbations markedly degrade unauthorized recovery. The approach enhances steganographic capacity and security for secure intellicise wireless networks, with practical implications for robust, private-content transmission in intelligent, constrained environments.

Abstract

Semantic Communication (SemCom), leveraging its significant advantages in transmission efficiency and reliability, has emerged as a core technology for constructing future intellicise (intelligent and concise) wireless networks. However, intelligent attacks represented by semantic eavesdropping pose severe challenges to the security of SemCom. To address this challenge, Semantic Steganographic Communication (SemSteCom) achieves ``invisible'' encryption by implicitly embedding private semantic information into cover modality carriers. The state-of-the-art study has further introduced generative diffusion models to directly generate stega images without relying on original cover images, effectively enhancing steganographic capacity. Nevertheless, the recovery process of private images is highly dependent on the guidance of private semantic keys, which may be inferred by intelligent eavesdroppers, thereby introducing new security threats. To address this issue, we propose an Agentic AI-driven SemSteCom (AgentSemSteCom) scheme, which includes semantic extraction, digital token controlled reference image generation, coverless steganography, semantic codec, and optional task-oriented enhancement modules. The proposed AgentSemSteCom scheme obviates the need for both cover images and private semantic keys, thereby boosting steganographic capacity while reinforcing transmission security. The simulation results on open-source datasets verify that, AgentSemSteCom achieves better transmission quality and higher security levels than the baseline scheme.

Figures (11)

• Figure 1: Network model of the proposed AgentSemSteCom scheme, including the agentic AI-based transmitter, physical channel, agentic AI-based receiver, key management center, and agentic AI-based eavesdropper. It is realized based on the edge-cloud library zhang2024space-air-ground and shared semantic knowledge base, with the aid of agentic AI's capabilities of environmental perception, task reasoning, and tool invocation.
• Figure 2: The proposed AgentSemSteCom scheme, where the designed five modules include the semantic extraction module, digital token controlled reference image generation module, conditional diffusion model-based coverless steganography module, JSCC-based semantic codec module, and optional task-oriented enhancement module.
• Figure 3: Visualization results of recovery images over different SNRs, where the public key is “a wild boar walking through a lush green field".
• Figure 4: Visualization results of AgentSemSteCom under different classes of images, which include common steganography images, facial images and style images. The images are collected at $\text{SNR = 10~dB}$.
• Figure 5: Comparison between AgentSemSteCom and SemSteDiff, which is simulated towards common steganography images.