AI & Tech Daily Briefing — 2026-09-13
1. South Korea broadens espionage law to protect chip technology
South Korea’s expanded espionage law took effect September 13, widening offenses beyond North Korea-related spying to cover activity benefiting any foreign country. The change is aimed partly at protecting strategically important semiconductor and other advanced technologies as global competition over chips intensifies.
Why it matters: South Korea is home to critical memory-chip suppliers, and the tougher law shows how semiconductor know-how is increasingly being treated as a national-security asset rather than ordinary corporate intellectual property.
Sources: Reuters, Financial Times
2. Trump pushes back on AI risk warnings as safety debate intensifies
U.S. President Donald Trump said concerns about artificial intelligence are being exaggerated and argued that the United States should avoid slowing development while it competes with China. His remarks came amid increasingly public warnings from leading AI researchers and executives about catastrophic-risk scenarios and calls for stronger safeguards.
Why it matters: The comments sharpen a central U.S. policy tension: whether aggressive AI development is essential for geopolitical leadership or whether frontier systems require stronger controls even at the cost of speed.
Source: Reuters
3. Anthropic CEO calls for coordinated slowdown in frontier AI development
Anthropic CEO Dario Amodei urged AI companies to slow capability advances while stronger safeguards are put in place, proposing independent evaluation, common safety standards among developers and greater international cooperation. The appeal followed Anthropic disclosures describing malicious uses of Claude, including cyber and biological-risk activity.
Why it matters: A leading frontier-model company publicly advocating slower development signals that safety concerns are moving from research circles into corporate strategy, potentially increasing pressure for industry-wide agreements and government regulation.
Sources: Reuters on Amodei’s proposal, Reuters on Anthropic’s threat report
4. Hyundai turns to Nvidia as its in-house autonomous-driving timeline slips
Hyundai Motor Group said vehicles using its internally developed driver-assistance software are now expected in late 2029, two years later than originally targeted. In the meantime, the South Korean automaker will work with Nvidia to deploy more advanced Level 2+ and Level 2++ driver-assistance systems by 2028 using Nvidia’s automotive platform.
Why it matters: The partnership illustrates how Nvidia’s AI stack is expanding beyond data centers into transportation, while Hyundai’s delay highlights the difficulty even major automakers face in building competitive autonomous-driving software independently.
Source: Reuters
5. South African opposition tests the limits of the global AI data-center boom
A proposed hyperscale Equinix data center near Cape Town International Airport is facing an appeal from campaigners who argue that planning rules have not adequately accounted for the project’s potential water and electricity demands. South Africa already hosts the majority of Africa’s data-center capacity, making the dispute an early test of how the continent will balance digital-infrastructure investment with local resource constraints.
Why it matters: AI growth depends on enormous physical infrastructure, and fights over electricity, water and permitting are becoming a global bottleneck. The Cape Town case shows that the AI infrastructure race is increasingly shaped by local environmental and regulatory limits, not just access to chips and capital.
Source: Financial Times