AI Upends Education, Google Nails Hurricane Forecast, Taco Bell Hits Pause

The Atlantic publishes a conversation between professor Ian Bogost and reporter Lila Shroff showing generative AI is now woven into daily work for students and teachers from high school through college. They note current seniors have never known academic life without ChatGPT, and AI-written essays and problem sets are routine.

Bogost says initial faculty panic has faded into a blind spot, while Shroff’s interviews reveal students use ChatGPT “all the time for everything”. She adds that three in ten K-12 teachers now use AI weekly and the College Board plans new AP Business and AP Cybersecurity courses to meet more practical demands.

Bogost believes the flood of AI has exposed years of 'pedagogical debt'—overloaded classes and outdated assignments that only worked before chatbots. Both writers describe AI as a breaking point that pushes schools to reconsider workload, teaching methods, and what counts as learning.

Taco Bell is reconsidering its voice AI ordering rollout after installing the system at more than 500 drive-throughs. Chief Digital and Technology Officer Dane Mathews says the technology will not be the default everywhere and human crews may reclaim busy lanes.

Glitches, delays, and prank requests such as “18,000 cups of water” have flooded social media, exposing reliability gaps. Even so, the AI has processed over two million orders and remains a “critical” item on parent Yum Brands’ roadmap.

Three years into the generative AI boom, fast-food drive-throughs remain stubbornly difficult for automation. Taco Bell’s rethink echoes McDonald’s scrapping an IBM pilot and Wendy’s cautious expansion with Google, underscoring industry-wide operational uncertainty.

Already profitable: 7x revenue growth, with 24% net margin (projected 30+% in 2026)

Huge audience: 5x bigger than CNN’s primetime viewership

Rapid growth: From 0 to 2.5m loyal readers in just two years 

MIT researchers launched VaxSeer, an AI system that predicts dominant influenza strains and recommends vaccine compositions months in advance. The open-access study detailing the tool appears in Nature Medicine.

In a 10-year retrospective analysis, VaxSeer’s picks outperformed the World Health Organization’s selections in nine of ten A/H3N2 seasons and matched or beat them in six of ten A/H1N1 seasons. Its predicted coverage scores also aligned with vaccine effectiveness data from the CDC, Canada’s Sentinel network, and Europe’s I-MOVE program.

The team says VaxSeer’s combined modeling of viral dominance and antigenicity cuts the guesswork that has long undermined flu vaccine planning. By mirroring real-world reductions in illness and medical visits, the system shows AI can narrow the costly lag between viral change and public health action.

Google's Weather Lab AI model delivered the most accurate 72-hour track and intensity forecasts for Hurricane Erin. Post-storm data show it bested the National Hurricane Center’s official forecast and every leading physics-based model.

Released in June and trained on reconstructed weather histories and specialized hurricane data, the model was already promoted as matching top systems. Erin validated that claim by also outperforming the trusted bias-corrected consensus models TVCN and IVCN.

Forecasters see the result as evidence that AI weather modeling now rivals the field’s stalwart physics tools after minimal development time. They say Weather Lab will carry more weight in future storm briefings, marking a rapid shift in operational trust.

San Mateo-based Alef struck agreements with Half Moon Bay and Hollister airports to operate its road-legal flying cars alongside conventional aircraft. Initial operations will include driving, vertical takeoff, forward flight, vertical landing, and both air and ground maneuvering.

The agreements raise Alef’s test locations to five, but this is the company’s first use of airports. Airport officials will maintain priority for regular aircraft while Alef assesses how its systems slot into established traffic patterns.

Alef says the move represents the first instance of cars and aircraft sharing the same operational space, as well as the first vertical takeoff of a car at an airport. The controlled environment gives the FAA, airport managers, and pilots a live view of the technology’s integration with existing procedures.