Technology

Governments are responding to the same supply-chain and infrastructure pressures. In Europe, the European Union is preparing a strategy to reduce technological dependence on the United States, with a focus on semiconductors, cloud computing and artificial intelligence. The plan includes a proposed Cloud and AI Development Act intended to accelerate construction of European data centers and streamline procedures, with a goal of tripling capacity within five to seven years.

The EU strategy would also require governments to conduct “sovereignty risk assessments” to identify vulnerabilities and strengthen domestic European alternatives. The approach signals that AI infrastructure is increasingly being treated as a matter of economic resilience, not merely private-sector innovation.

In the United States, state-level regulation is advancing alongside the infrastructure race. Connecticut Governor Ned Lamont signed the Online Safety Act, Public Act No. 26-15, creating a broad framework covering artificial intelligence and online platforms. The law takes effect October 1, 2026, and includes provisions for consumer-facing AI subscriptions, AI companions and automated employment-related decision technology.

The Connecticut law requires risk detection protocols and restrictions involving AI companions’ interactions with minors. It also requires transparency and notice for applicants and employees when automated employment-related decision technology is used. The measure positions Connecticut among states moving to define rules for AI systems, following similar legislative activity in Illinois and reflecting growing interest in oversight of frontier AI models and consumer-facing applications.

Safety scrutiny is also intensifying around AI-enabled products already on the market. A Reuters investigation found that Tesla AI data labelers and a former self-driving engineer expressed distrust in the company’s “Full Self-Driving” system and in the safety statistics used to promote it. The report, based on interviews with former employees and traffic-safety researchers, said the methodology behind Tesla’s safety claims was flawed.

According to the investigation, data labelers regularly observed Full Self-Driving failures in basic driving tasks, including responding to emergency vehicles, navigating construction zones and recognizing pedestrians. Some former employees said they would not trust the technology to drive them.

Taken together, the day’s developments show the AI boom entering a more complex phase. Companies are racing to secure the chips and data center capacity needed to deploy new products, investors are rewarding semiconductor suppliers tied to that buildout, and governments are seeking greater control over the infrastructure and rules that will shape the market. At the same time, investigations into deployed AI systems are highlighting the gap between ambition and verified reliability.