美国海关与边境保护局正在推进一项以“量子传感器”为核心、结合人工智能数据库的缉毒计划，目标是在车辆和集装箱中更有效识别芬太尼等毒品。根据联邦公报披露的合同说明，CBP已为此向通用动力支付240万美元，该合同于2025年12月公开。项目背景是国土安全部推动AI规模化应用的整体战略，且CBP在2025年4月至10月期间完成了市场调研，显示边境查获阿片类毒品数量持续上升。
从现有线索看，CBP此前在2025年7月曾寻求采购35台手持式“Gemini”化学分析仪，用于识别芬太尼、氯胺酮、可卡因、甲基苯丙胺和MDMA等物质。这类设备依赖傅里叶变换红外光谱和拉曼光谱技术，DHS在2021年和2023年已进行过测试。但研究指出，便携式光谱仪在芬太尼检测中可能存在误报与漏报风险。合同说明称，CBP评估过至少十余种设备，其中一种“无法检测芬太尼”，这也成为开发新型传感方案的动因。
所谓“量子传感器”可能借鉴量子化学方法。2024年的一项研究显示，利用量子点与荧光染料，可检测芬太尼及其58种类似物，灵敏度达到微克级。人工智能在其中的潜在作用是处理复杂混合信号，通过光谱解卷积识别人眼难以区分的成分。专家指出，AI结合新型传感材料，或能在街头毒品快速变化的情况下，提高检测准确率，但从实验室到边境现场仍存在巨大技术与实施挑战。
US Customs and Border Protection is pursuing an initiative to detect fentanyl using so-called “quantum sensors” paired with an artificial intelligence database. A contract justification published in the federal register shows CBP is paying General Dynamics $2.4 million for a prototype, a contract made public in December 2025. The effort aligns with a broader Department of Homeland Security push to scale AI adoption, following CBP market research conducted from April through October 2025 amid rising opioid seizures at US borders.
Available details suggest CBP previously explored conventional tools. In July 2025, the agency sought 35 handheld “Gemini” analyzers designed to identify drugs such as fentanyl, ketamine, cocaine, methamphetamine, and MDMA using Fourier Transform Infrared and Raman spectroscopy. DHS tested similar devices in 2021 and 2023, but research indicates portable spectrometers can struggle with fentanyl detection, producing false positives or negatives. The justification notes CBP evaluated more than ten devices and found at least one that “cannot detect fentanyl,” motivating interest in alternative approaches.
The “quantum” concept may draw on quantum chemistry techniques. A 2024 study demonstrated that quantum dots combined with fluorescent dye can detect fentanyl and 58 analogues at microgram levels. AI could assist by performing spectral deconvolution, separating overlapping signals in complex mixtures that humans might miss. Experts caution that such methods are typically validated only in tightly controlled lab settings, suggesting that while AI-enhanced quantum sensing could improve accuracy, translating it to real-world border inspections presents significant technical hurdles.