Rest of World reports that machine translation in Tigrinya rendered smallpox as syphilis, gonorrhea as diabetes and “you have been given intravenous antibiotics” as “you have been given intravenous insecticides”, and that more than two-thirds of chatbots do not adequately account for dialects or recognise urgency cues, according to a review in India.
Colin Ashruf, an independent advisor, reports that much of the language is imprecise, anthropomorphic, and misleading, and that calling something intelligent, even artificially intelligent, implies it is more than just a thing or a tool.
Semafor reports that most participants thought China could rely on its economic and technological engine to ride out storms, and that China’s immense scale, national unity, deep political control, and increasingly cutting-edge technology may allow it to sustain superpower status even if it fails to resolve long-standing structural difficulties.
The authors of “AI Agents Push Humans Out of the Loop” argue that current approaches to AI agent design impede effective human oversight and also degrade the cognitive capacities required for it.
Meeting the Coming Wave reports that compensation accounts for 2.3% of response-frame mentions, enablement and investment dominate at 55.2%, regulation and restriction follow at 21.8%, and training at 20.6%, and that the mainstream and radical right support unrestricted enablement while the left is critical but divided on remedy.
State Comptroller Tom DiNapoli found that New York City has only partially implemented 2023 recommendations on government AI use, lacking a complete inventory of AI tools, consistent citywide standards, and a formal process for public concerns.
The Austrian government plans to have official notices automatically generated by language models, with human oversight only through spot checks.
Restrepo stated that only 35% of state procedures are digitalized and that Colombia generates around 7,000 new norms, laws, decrees, resolutions or circulars per year, equivalent to about 20 per day.
The Jagged Global Economy: Frontier AI Unevenly Exposes National Economies reports that high-income countries are substantially more exposed than low-income countries, Europe and Central Asia are 50 percent more exposed than Sub-Saharan Africa, and women are more exposed than men in 91 percent of countries, driven by their concentration in white-collar and sales occupations.
Today’s lead: The score can be the shortcut.