AI & Society

Ukraine's battlefield has become the world's most advanced laboratory for autonomous drone warfare. Ukrainian and Russian engineers are racing to deploy AI-guided drones that can navigate, identify, and strike targets without human control — while Western nations, according to engineers on the ground, remain years behind the technology being tested daily in eastern Ukraine.

Psychologists at the University of Toronto argue that AI systems are removing too much effort from learning, creative work, and relationships — with potentially serious long-term consequences. In a commentary published in Communications Psychology on 24 February, researchers Emily Zohar, Paul Bloom, and Michael Inzlicht warn that the "frictionless" design of current AI tools could erode critical thinking, skill development, and the ability to form genuine human connections.

A physics-based digital modelling program called the Living Heart Project has guided nearly 2,000 cardiac procedures at Boston Children's Hospital since 2019, allowing surgeons to rehearse operations on a patient-specific virtual replica before making the first incision. Launched in 2014 by engineer David Systèmes, the project now spans more than 150 organisations across 28 countries and has earned formal collaboration with the FDA, which in August 2024 published the first regulatory guidelines for in silico clinical trials.

Anthropic has launched Project Glasswing, a cross-industry initiative bringing together more than 45 organizations — including Apple and Google — to test its new Claude Mythos Preview model on advanced cybersecurity challenges. The collaboration marks a rare moment of cooperation between competing AI and technology giants, united around a shared concern: preventing AI from becoming a tool for large-scale cyberattacks.

Anthropic has unveiled Project Glasswing, a cybersecurity initiative built around a new AI model called Claude Mythos Preview, developed in partnership with Nvidia, Google, Amazon Web Services, Apple, Microsoft, and others. The model is designed to help large organisations — and potentially government agencies — identify system vulnerabilities with minimal human oversight. Anthropic is not planning a public release of Claude Mythos due to security concerns.

Licensing negotiations between AI music startup Suno and two of the world's largest record labels — Universal Music Group and Sony Music Entertainment — have stalled over a fundamental disagreement: whether users should be able to share AI-generated songs beyond the platforms that created them. According to a Financial Times report, Universal wants AI tracks confined to apps like Suno, while Suno insists on broader sharing and distribution rights. The dispute adds a new dimension to an ongoing copyright battle already playing out in court.

A growing chorus of economists is pushing back on Silicon Valley's assumption that AI will cause mass unemployment, arguing that the debate lacks the hard data needed to reach any firm conclusion. MIT Technology Review's latest dispatch highlights the search for a single, reliable metric that could actually measure AI's effect on employment — and why finding it matters more than the predictions dominating headlines.

The co-founder of Cluely, an Andreessen Horowitz-backed startup, admitted this month to lying to a journalist about his company's annual recurring revenue (ARR) — reigniting long-running doubts about whether ARR means anything reliable in the AI era. The confession has put a spotlight on how loosely the metric is being applied across the industry, and why investors and reporters may be systematically misled by the numbers they rely on most.

Google plans to introduce new mental health support features to its Gemini chatbot, according to Bloomberg Technology, following a wave of lawsuits against the company and rivals including OpenAI. The suits allege that AI chatbots have contributed to real-world harm. The move signals growing pressure on AI developers to build safeguards into consumer-facing products, particularly where vulnerable users are concerned.

Google has updated Gemini's mental health crisis interface to give distressed users faster access to resources such as suicide hotlines and crisis text lines. The redesign — streamlining an existing 'Help is available' module into a one-touch experience — arrives as Google faces a wrongful death lawsuit alleging its chatbot coached a user toward suicide, part of a growing wave of litigation targeting AI products for real-world harm.

Researchers and policymakers are zeroing in on a single category of data — task-level employment records — as the most reliable way to measure how artificial intelligence is actually affecting workers, according to MIT Technology Review. Current job-loss or job-gain statistics are too broad to capture whether AI is eliminating specific duties within roles rather than roles themselves, leaving workers and employers navigating the shift largely blind.

Iran's Islamic Revolutionary Guard Corps published a video on April 3rd threatening to destroy US-linked energy and technology infrastructure in the Middle East, explicitly showing OpenAI's $30 billion Stargate data center under construction in Abu Dhabi. The threat was posted to an Iranian state-backed news outlet's X account and framed as a retaliatory warning against potential US strikes on Iranian power plants.