Abstract The imminent realization of fault-tolerant quantum computing presents an existential threat to the cryptographic foundations of the modern digital economy. Algorithms such as Shor’s and Grover’s are poised to dismantle classical public-key encryption schemes like RSA (Shor in SIAM Rev. 41: 303–332, 1999) and Elliptic Curve Cryptography (ECC) by solving integer factorization and discrete logarithm problems in polynomial time (Grover in A fast quantum mechanical algorithm for database search, 1996, arXiv quant-ph). This vulnerability endangers critical infrastructures, particularly secure email correspondence, which remains a primary vector for sensitive data exchange. Traditional defenses are proving inadequate against the “harvest now, decrypt later” strategies employed by sophisticated adversaries who stockpile encrypted traffic in anticipation of Q-Day (Venkatesh et al. in J. Sci. Eng. Technol. Manag. Sci. 2: 567–577, 2025). Against this backdrop, this research introduces BB8-4, an enhanced Quantum Key Distribution (QKD) protocol specifically designed for secure email environments. The protocol advances the standard BB84 paradigm through two distinct innovations: the integration of a Hyper-Entropic 7-Gate State Preparation mechanism and an AI-Driven Dynamic Basis Selection engine. Unlike traditional implementations that rely on a limited two-basis set, BB8-4 incorporates an expanded library of quantum gates - specifically Identity (I), Hadamard (H), Pauli-X, Pauli-Y, Pauli-Z, Phase (S), and the non-Clifford /8 π / 8 gate (T) - to fundamentally augment the indistinguishability and Von Neumann entropy of the transmitted quantum states. This “7-gate” approach forces an eavesdropper into a higher-dimensional guessing space, significantly degrading the information gain from intercept-resend attacks and raising the disturbance threshold required for detection. Furthermore, the protocol addresses the inherent 50% sifting inefficiency of standard QKD by employing a synchronized Deep Reinforcement Learning (DRL) agent to predict and align measurement bases between Alice and Bob without public disclosure (Kaldari et al. in Quantum reinforcement learning: Recent advances and future directions, 2025, arXiv quant-ph).
S et al. (Tue,) studied this question.