Alphabet clarifies robotaxis aren’t remotely driven amid oversight; Google warns of cyberthreats, AI chip shifts

Alphabet clarifies robotaxi operations as lawmakers press on oversight

Alphabet’s self‑driving unit says its robotaxis are never remotely driven in U.S. operations, a public clarification aimed at easing mounting legislative and public concerns about safety and offshore intervention. The company stresses that steering, acceleration and braking decisions on U.S. roads are made by onboard systems, and that so‑called remote intervention — which may involve technical support staff located abroad — does not amount to real‑time vehicle control during passenger trips. Alphabet frames the statement as a corrective to suggestions that cars could be piloted from afar and to worries that overseas teams might influence on‑road behaviour.

The disclosure directly addresses queries from lawmakers who have flagged potential safety, accountability and transparency risks if vehicles could be operated remotely or if control functions were effectively offshored. Alphabet emphasises that overseas staff perform ancillary technical roles and support tasks rather than driving the vehicle, and says its operational model for domestic deployments keeps human oversight distinct from active vehicle control. The company presents the clarification as part of broader efforts to inform legislative inquiries and to provide regulators, riders and the public with clear descriptions of how onboard autonomy, safety protocols and human oversight interact in everyday service.

Alphabet also signals the move is designed to reduce regulatory friction while it scales services, by spelling out practices that separate diagnostic or supervisory support from direct driving. Company officials seek to reassure policymakers that U.S. robotaxi services remain governed by automated onboard decision‑making and established safety layers, even as they acknowledge the operational role of global engineering teams in maintenance, software updates and non‑driving support functions.

Google cybersecurity analysis highlights state‑linked threats

Separately, Google Threat Intelligence reports that the U.S. defence industrial base is facing sustained cyber intrusions from groups tied to China, Russia and North Korea, with China‑nexus actors accounting for the most activity. The analysis warns of tactics such as edge‑device exploitation, reconnaissance via ORB networks and use of large AI models for open‑source intelligence, and calls for stronger monitoring and cross‑sector information‑sharing to protect aerospace and defence suppliers.

AI hardware ties shift industry access dynamics

Industry observers note that intensified hyperscaler‑supplier partnerships are reshaping access to high‑performance AI chips. An expanded long‑term deal between a major cloud operator and Nvidia is cited as an example of how preferential supply, co‑development and technical prioritisation can raise barriers for rivals seeking equivalent accelerator hardware and software optimisation.