AI Models Reveal Stark Forecasts for Middle East Conflict

Decrypt asked seven top AI models to forecast the Israel–Iran confrontation. Six predicted a drawn-out campaign of airstrikes, cyberattacks, and proxy clashes, while ChatGPT alone anticipates a near-term diplomatic deal.

Across the six pessimistic forecasts, the models cite mutual deterrence, U.S. reluctance to intervene, and survival calculations as brakes on escalation. They foresee intermittent conflict lasting from three to 24 months, with risks tied to nuclear timelines and rogue proxies.

The article notes the shared verdict amounts to a “shadow war in daylight”—more frequent strikes without tipping into open war. This uniform outlook highlights how the models converge on survival-driven deterrence as the defining constraint for both capitals.

Starbucks unveiled Green Dot Assist, a generative-AI virtual assistant built to help baristas with instant, conversational answers during shifts. The system sits on in-store iPads and replaces manual searches for beverage recipes and operational guidance.

The company is piloting the tool in 35 coffeehouses and says more locations are on the way. Starbucks frames Green Dot Assist as a move to streamline operations and reduce friction for its frontline workers.

Immediate access to accurate information lets baristas stay focused on crafting drinks and engaging customers. Leadership calls it the start of a new era in which AI elevates the partner experience and deepens customer connection.

BT chief executive Allison Kirkby said advances in artificial intelligence could push the company to reduce its workforce beyond the 55,000 job cuts already announced. She stated BT may be “even smaller” by 2030 if AI efficiencies materialise.

BT’s current plan targets 40,000–55,000 redundancies and £3bn in cost savings by decade’s end. Kirkby has already sold Italian and Irish units and carved out the international division to tighten focus on the UK market.

The remarks tie BT’s future headcount directly to AI adoption rather than traditional restructuring levers. Kirkby also signalled that if the market keeps undervaluing Openreach, BT will have to revisit a potential separation once the fibre upgrade is complete.

A Guardian investigation shows UK universities recorded nearly 7,000 confirmed cases of students cheating with ChatGPT and similar tools in 2023-24. The rate jumped to 5.1 incidents per 1,000 students, triple the previous year’s 1.6.

Early data suggest the current academic year is tracking toward 7.5 AI cheating cases per 1,000 students. Traditional plagiarism cases have fallen from 19 to 15.2 per 1,000 over the same period, and more than a quarter of universities still do not log AI misuse as a distinct offence.

Researchers who covertly submitted AI-generated assignments reported a 94% success rate in evading detection. Academics interviewed call the confirmed violations “the tip of the iceberg,” underscoring how current assessment and monitoring systems are struggling to keep pace with generative AI.

An Instagram reel from account @daily_loud shows a man passionately kissing a lifelike humanoid robot. The clip has gone viral, drawing wide attention across social media.

The robot blinks, appears to reciprocate the kiss, and displays facial reactions that blur the line between human and machine. Commenters express astonishment and alarm, highlighting how convincingly the scene mimics genuine intimacy.

The episode spotlights growing public unease about forming emotional bonds with AI and potentially neglecting human relationships. It also underscores broader security concerns as advances in realism make it harder to distinguish real images from digital creations.