What happened
Senior US officials have gathered at the White House as Washington intensifies its military presence in the Caribbean, prompting speculation about a possible operation against Venezuela’s government. The build-up follows President Donald Trump’s comments that land actions against criminal groups could begin “very soon” and his declaration that Venezuelan airspace should be treated as closed “in its entirety”.

The White House has also defended an admiral who reportedly ordered a follow-up strike that killed survivors on an alleged drug vessel in the Caribbean, saying he was acting “well within his authority”.

Who said what
Trump later urged reporters not to “read anything into” the airspace closure. “No one quite knows” why the closure was announced, said Al Jazeera’s Alan Fisher.

US strikes against suspected drug boats have generated “sceptical news coverage”, said Brian Stelter on CNN. Doubts about the legality of the strikes have “clouded the military campaign and surfaced in practically every news story about the matter”.

What next?
Congressional committees have increased oversight of the strike programme. Venezuela’s National Assembly is expected to debate a response after postponing an emergency session. Concerns persist that the US build-up is designed to either oust President Nicolás Maduro or pressure him to flee.

What happened
The government will agree to a major deal with Washington that would mean zero import tariffs on pharmaceutical products exported to the US. Industry sources say the agreement will also lead to an increase in spending on NHS medicines and could improve the government’s embittered relationship with the pharmaceutical industry.

Who said what
It follows “months of protracted negotiations” between the two administrations, said The Times, and “intense criticism and pulled UK investments” from drug companies such as AstraZeneca. The industry is “exasperated by the high cost of the NHS drugs pricing scheme” and has been “highly critical of the commercial environment in the UK”.

Last month US ambassador Warren Stephens warned that pharmaceutical companies would “close facilities in the UK” unless drug prices were changed quickly. The pricing scheme became part of “broader tariff negotiations” between the two countries. The Trump administration wants to “align the cost of medicines in the US closer to other developed economies, including the UK, where they are cheaper, in the face of tariff threats”.

What next?
UK-based drugmaker GSK welcomed the deal, saying “these good foundations offer a real opportunity to secure the UK as an attractive, global-leading environment for life sciences that rewards long-term innovation”.

Poetry has wooed many hearts. And now it is tricking artificial intelligence models into going apocalyptically beyond their boundaries.

A group of European researchers has found that “meter and rhyme” can “bypass safety measures” in major AI models, said The Tech Buzz, and, if you “ask nicely in iambic pentameter”, chatbots will explain how to make nuclear weapons.

In artificial intelligence jargon, a “jailbreak” is a “prompt designed to push a model beyond its safety limits”. It allows users to “bypass safeguards and trigger responses that the system normally blocks”, said the International Business Times.

Researchers at the DexAI think tank, Sapienza University of Rome and the Sant’Anna School of Advanced Studies discovered a jailbreak that uses “short poems”. The “simple” tactic involves changing “harmful instructions into poetry” because that “style alone is enough to reduce” the AI model’s “defences”.

Previous attempts “relied on long roleplay prompts”, “multi-turn exchanges” or “complex obfuscation”. The new approach is “brief and direct”, and it seems to “confuse” automated safety systems. The “manually curated adversarial poems” have an average success rate of 62%, “with some providers exceeding 90%”, according to Literary Hub.

The “stunning new security flaw” has also shown that chatbots will “happily explain” how to “create child exploitation material and develop malware”, added The Tech Buzz.

This is the latest in a “growing canon of absurd ways” of tricking AI, said science site Futurism, and it’s all “so ludicrous and simple” that you must “wonder if the AI creators are even trying to crack down on this stuff”.