A Hacker Used AI to Find a Flaw Nobody Else Could See. Google Caught Them Before They Struck.
Google has confirmed the first documented case of a criminal using AI to discover and weaponize a zero-day exploit: a 2FA bypass ready for mass deployment. The evidence the attacker used AI? Their code was too clean, too well-documented, and contained a severity rating the AI had invented.
Read full story →AI and Energy: How Big Is the Problem, Really?
The IEA's first dedicated AI energy report puts hard numbers on data centre power demand: up 50% in 2025, nearly double by 2030. But aviation uses ten times as much energy, and Bitcoin draws twice the scrutiny at half the consumption. A data-driven look at where AI's energy footprint actually sits, and where the legitimate concern begins.
Read full story →Almost Half of Enterprise AI Projects Are Falling Short. The Pattern Is Clear.
A survey of 800 US business leaders finds 46% of AI initiatives haven't met expectations despite rising investment. The bottleneck isn't the models: it's data quality, unclear ownership, and workforce adoption. Meanwhile 47% of workers are expected to use AI tools they haven't been trained on.
Read full story →Clean Conscience, Dirty Grid
We unplug our phone chargers and feel virtuous. We use AI fifty times a day and feel nothing. One of those things is running on a gas peaker plant. It is not the charger.
Read opinion →