AI agents, export curbs and deepfake ads test tech's next phase

AI systems that can execute multi-step tasks without constant user input are moving from prototypes into consumer devices, while U.S. export controls on advanced models and widespread use of synthetic media in campaigns create new points of friction. Qualcomm CEO Cristiano Amon told CNBC that the company is preparing more than 40 designs for wearable products, including jewelry, camera-equipped earbuds and pins, that would keep an agent in constant contact with the user and the surrounding environment. He described these agents as successors to current apps, able to retrieve banking details or book travel across services without manual navigation. Amon also projected that smart-glasses shipments could reach hundreds of millions of units within a couple of years, rivaling smartphone volumes recorded at 1.26 billion in 2025 by Counterpoint Research. Separate administration actions placed Anthropic's Mythos 5 and Fable 5 models under export controls that restrict access by foreign nationals, citing a bypass technique affecting cybersecurity safeguards. Canadian Prime Minister Mark Carney and European Commission Executive Vice-President Henna Virkkunen publicly questioned reliance on U.S. models in response. In parallel, multiple 2026 primary campaigns deployed AI-generated video and images, including a Texas Senate ad depicting Democratic nominee James Talarico in fabricated scenarios and a Kentucky district contest featuring deepfakes of Rep. Thomas Massie with other lawmakers. Campaigns on both sides used the tools, and no federal disclosure rule currently requires labeling. The three developments share an unresolved tension: whether rapid capability gains in agents and hardware can coexist with stable access guarantees for allies and verifiable standards for political content.