Immunotherapy breakthrough for bowel cancer & China’s rapid everyday AI adoption - News (May 6, 2026)

We’ll start in health, with results that are turning heads in oncology. A UK-led clinical trial tested the immunotherapy drug pembrolizumab before surgery in a specific subtype of stage II to III colorectal cancer—tumors described as MMR-deficient or MSI-high. Early checks suggested many patients had no detectable cancer after treatment and surgery, but the longer follow-up is what really stands out: after a median of about 33 months, none of the participants had relapsed. If this holds up in larger studies, it could change the usual playbook—shifting some patients away from months of post-surgery chemotherapy and toward a faster, more targeted start. Researchers also highlighted blood-based signals that may help predict who’s responding, which could make treatment more personalized and less disruptive.

Now to AI—and to a place that’s becoming a real-world stress test for how quickly these systems can blend into daily life. In China, generative and so-called “agentic” AI is spreading beyond tech demos and into routine tasks, from trip planning and food orders to hiring workflows and health monitoring. The article describes crowds in major cities looking for help setting up tools, a sign that adoption is moving from early adopters to the mainstream. Government data puts China at more than 600 million generative-AI users as of December, a steep jump from a year earlier. What’s especially notable: usage intensity is rising fast. Data cited from OpenRouter suggests Chinese models are now consuming more tokens each week than US models—meaning more day-to-day prompts, more automated actions, and more “learning by doing” at scale. Big companies like Tencent, Alibaba, and Baidu are weaving AI into the platforms people already use, shifting the race from standalone chatbots to full ecosystems. Export limits on advanced chips remain a constraint, but analysts argue those limits may also be pushing faster domestic coordination—sometimes with support from Huawei-linked hardware. The broader question: as AI becomes normalized inside China’s tightly managed internet, the balance it strikes between speed, usefulness, and control could shape global expectations for adoption and governance.

In the US, a quieter but important move on AI oversight: Google, Microsoft, and Elon Musk’s xAI say they’ll voluntarily submit new models for safety testing before public release. The evaluations will run through the Commerce Department’s Center for AI Standards and Innovation—CAISI—and focus on what the models can do, how secure they are, and what risks they might pose to public safety or national security. This builds on earlier voluntary testing frameworks, but the timing is telling. Even as Washington debates how heavy a regulatory hand to use, the pressure is rising—especially with AI increasingly tied to defense, intelligence, and critical infrastructure. Voluntary testing isn’t the same as binding regulation, but it does signal a growing expectation: powerful models should be scrutinized before they hit the open market.

Two major AI court fights are also sharpening how the industry may be governed—by contract, by copyright, and by precedent. First, in a civil trial tied to OpenAI’s origins and its later shift toward a high-value, profit-oriented structure, OpenAI President Greg Brockman testified that his stake is worth nearly 30 billion dollars—despite saying he didn’t personally invest money into the company. The valuation cited in court, about 852 billion dollars, underlines why this dispute is so explosive. At the center is the claim that OpenAI’s leadership betrayed the organization’s early mission and expectations set with early backers, including Elon Musk. The judge also kept certain alleged text messages out of evidence, a reminder that courtroom battles often hinge as much on procedure as on drama. The bigger point: as AI labs grow into financial giants, their founding promises and governance structures are no longer philosophical—they’re legal and monetary fault lines.

Second, a new class-action lawsuit in Manhattan pairs bestselling author Scott Turow with major publishers including Hachette, Macmillan, McGraw Hill, Elsevier, and Cengage. They accuse Meta—and Mark Zuckerberg—of training Llama models on massive volumes of copyrighted books and journal articles allegedly sourced from pirate libraries. Meta denies wrongdoing and argues that AI training can qualify as fair use, but the plaintiffs are pushing for damages, an injunction, and destruction of the disputed copies. What makes this case particularly pivotal is the piracy allegation: courts may view “transformative use” differently when the starting point is unlicensed material from known shadow libraries rather than negotiated datasets.

Let’s talk markets, where AI enthusiasm is still moving trillions of dollars around—just with a shifting cast of winners. Alphabet is closing in on Nvidia in market value, driven by its AI strategy and a surge in Google Cloud growth. Investors appear to be rewarding companies that can turn AI into recurring enterprise revenue—not just sell the picks and shovels. Alphabet has also gained credibility as an AI chip player through its custom processors, and strong cloud numbers are reinforcing the story that its heavy spending is translating into real demand. Meanwhile, Micron jumped after announcing it’s shipping a record high-capacity data-center solid-state drive, and the broader rally reflects a key reality of the AI boom: memory—both DRAM and NAND—has become a bottleneck. Training and running large models isn’t only about GPUs; it’s also about storing and feeding enormous volumes of data quickly and efficiently. As shortages persist, memory suppliers are capturing a bigger share of AI’s economics. And ahead of today’s open, US stock futures rose on a separate catalyst: a report that the US believes it’s close to an agreement with Iran to end the war. That lifted “risk-on” sectors like travel while weighing on some energy-linked names. On top of that, chip-and-server stocks got a jolt from strong results and outlooks from companies like AMD and Super Micro, suggesting demand for AI infrastructure remains resilient.

Now to geopolitics, where the stakes are measured in both lives and oil shipments. China’s foreign minister Wang Yi is calling for an urgent, comprehensive ceasefire in the two-month war between the United States and Iran, after meeting Iran’s foreign minister in Beijing. The diplomatic push matters because of the Strait of Hormuz—a narrow passage that’s critical for global energy flows. Disruptions and delays there have rattled markets, even as prices fluctuate day to day. In Washington, President Donald Trump said he was pausing a US effort to guide stranded commercial ships through the strait, pointing to progress toward an agreement and requests from other countries. But reports also describe a fragile reality on the water: intermittent fighting, limited transit routes, and hundreds of merchant vessels still caught in limbo. The talks are also happening in the shadow of a planned Trump visit to Beijing for a summit with President Xi Jinping—adding diplomatic urgency and a high-profile deadline.

Finally, a significant shift in the South Caucasus: Armenia has hosted its first bilateral summit with the European Union, a symbolic and practical step as Yerevan pursues closer integration with Europe and loosens its long reliance on Russia. The sides signed a connectivity partnership aimed at strengthening transport, energy, and digital links, alongside deeper security cooperation. This pivot has a clear backdrop: Armenia’s trust in Moscow eroded after Azerbaijan retook Nagorno-Karabakh in 2023, and Yerevan has since taken steps that distance it from Russia, including freezing participation in the CSTO and joining the International Criminal Court. Armenia still faces trade-offs, given its economic ties in Russia-led structures, and the move also intersects with EU-Azerbaijan tensions over prisoners and rights concerns after the Karabakh exodus. In short, Armenia is widening its options—but in a region where every new partnership can create new friction.