1. Groundbreaking Study Finds AI Judges Are Too Easily Persuaded
Legal Futures · Law & Legal
A landmark study by academics at Maynooth University and University College Dublin found that AI models acting as judges were dangerously susceptible to persuasion by skilled advocates, potentially threatening fair outcomes in courts. Tested across 20 configurations of major AI models, every system showed measurable persuadability — meaning a better-argued case, not necessarily the legally correct one, could sway a verdict. Researchers warned this raises serious fairness concerns and called for persuadability to become a standard benchmark before any AI system is deployed in courts or tribunals.
2. AstraZeneca CEO: AI Boosts Drug Trial Success Rates and R&D Productivity
Healthcare Digital · Healthcare & Pharmaceutical
AstraZeneca CEO Pascal Soriot says AI is delivering tangible productivity gains across the pharma industry, helping companies design better drugs faster and make smarter decisions about which candidates to advance. Through its partnership with Tempus AI, AstraZeneca has built an AI agent that combines clinical and lab data to predict the probability of success in Phase 3 trials — potentially saving hundreds of millions of dollars per trial. Soriot also highlighted AI’s role in removing potential side effects from molecules during the design stage, signalling a broader shift in how big pharma is embedding AI into core R&D workflows.
3. AI Citations Emerge as Key Visibility Metric for Publishers and Brands
BestMediaInfo · Media & Marketing
Comscore’s Custom IQ head says AI citations are rapidly becoming a visibility and trust metric for publishers and brands, as consumer journeys shift beyond traditional search and clicks. As AI-powered answer engines reshape how audiences discover content, the ability to be cited by AI systems is emerging as a critical new measure of digital reach and credibility for both media owners and marketers.
4. Retired Army General Warns AI and Quantum Computing Threaten National Security
Military.com · Defense & Security
Retired Army Lt. Gen. Ross Coffman, president of Forward Edge-AI, warned on June 16 that while AI is rapidly transforming military operations and making cybersecurity threats more sophisticated, the looming arrival of practical quantum computing poses an equally significant — and underappreciated — danger to national security. He explained that governments and private industry are beginning to prepare for a future in which quantum computers could undermine many of today’s encryption systems, putting classified communications and defence infrastructure at risk.
5. Microsoft Raises 2026 AI CapEx to $190B Amid Component Price Surges
The Register · Technology & AI
Microsoft has announced it expects total 2026 capital expenditure to reach $190 billion, with $25 billion of the increase driven by surging memory and storage component prices tied to AI infrastructure demand. Despite spending $97 billion over the last four quarters, Microsoft’s AI services have generated $37 billion in annual recurring revenue, raising ongoing return-on-investment concerns on Wall Street. The scale of the spending underscores how the race to build AI infrastructure is pushing costs to unprecedented levels across the industry.
6. NVIDIA Shows AI Agents Cut City Incident Response Times by 80%
NVIDIA Blog · Construction & Infrastructure
NVIDIA has highlighted how smart city AI agents are delivering measurable results at scale, with Kaohsiung City in Taiwan cutting incident response times by 80% using street-level AI, and Raleigh, North Carolina, achieving 95% vehicle detection accuracy. French rail networks are also using AI-powered digital twins to optimise energy consumption by 20%. The results underscore how cities are moving from AI pilots to operational deployments that directly improve urban infrastructure performance.