Why California Is Swinging Right on Crime | Ransomware Is ‘More Brutal’ Than Ever in 2024 | How Game Theory Can Make AI More Reliable, and more

But then he promptly released an AI policy road map guaranteed to widen that gulf. In May, the Schumer-led Bipartisan Senate AI Working Group released a report after a nearly yearlong process of educational briefings, nine “Insight Forums,” and engagement with more than 150 experts and produced a proposal with few specifics other than urging the federal government to spend $32 billion a year on (nondefense) AI innovation.
The task of formulating actual regulations and policy designed to build public trust in AI was pushed off to unnamed “relevant committees.”
Even on matters as weighty as national security, the road map concluded with a grab bag of disjointed advice. Overall, it appears that Congress is worried about throwing sand in the gears of the AI industry’s innovation agenda. Its operating mantra appears to be: move slowly and make sure not to break anything.

Ransomware Is ‘More Brutal’ Than Ever in 2024  (Jordan Pearson, Wired)
Today, people around the world will head to school, doctor’s appointments, and pharmacies, only to be told, “Sorry, our computer systems are down.” The frequent culprit is a cybercrime gang operating on the other side of the world, demanding payment for system access or the safe return of stolen data.
The ransomware epidemic shows no signs of slowing down in 2024—despite increasing police crackdowns—and experts worry that it could soon enter a more violent phase.
“We’re definitely not winning the fight against ransomware right now,” Allan Liska, a threat intelligence analyst at Recorded Future, tells WIRED.

How Game Theory Can Make AI More Reliable  (Steve Nadis, Wired)
Imagine you had a friend who gave different answers to the same question, depending on how you asked it. “What’s the capital of Peru?” would get one answer, and “Is Lima the capital of Peru?” would get another. You’d probably be a little worried about your friend’s mental faculties, and you’d almost certainly find it hard to trust any answer they gave.
That’s exactly what’s happening with many large language models (LLMs), the ultra-powerful machine learning tools that power ChatGPT and other marvels of artificial intelligence. A generative question, which is open-ended, yields one answer, and a discriminative question, which involves having to choose between options, often yields a different one. “There is a disconnect when the same question is phrased differently,” said Athul Paul Jacob, a doctoral student at the Massachusetts Institute of Technology.
To make a language model’s answers more consistent—and make the model more reliable overall—Jacob and his colleagues devised a game where the model’s two modes are driven toward finding an answer they can agree on. Dubbed the consensus game, this simple procedure pits an LLM against itself, using the tools of game theory to improve the model’s accuracy and internal consistency.

Hackers Impersonating As Fake Toll Payment Processor Across The U.S.  (Tushar Subhra Dutta, Cyber Security News)
A convincing phishing scam recently targeted a cybersecurity researcher’s (Jonathan Munshaw) wife with a fake text claiming she owed unpaid New Jersey E-ZPass highway tolls.
The message prompted visiting a fraudulent site mimicking the legitimate E-ZPass website to pay fictitious fines. 
The timing coincided with a recent trip to the state where they have active E-ZPass accounts for automatic toll payment, lending credibility to the scam attempt. 
E-ZPass is an electronic toll collection system used across multiple U.S. states and managed by various agencies, which the scammers exploited by spoofing New Jersey’s official website alarmingly well.
Cybersecurity researchers at Talos Intelligence recently discovered that hackers had been actively impersonating fake toll payment processors across the U.S.

The Race to Get Ahead One of the Deadliest Natural Disasters  (Umair Irfan, Vox)
If you felt the ground starting to shake, would you know what you’re supposed to do?
You should. You might be more vulnerable to an earthquake than you realize: According to the US Geological Survey, almost 75 percent of the US population lives in an area that could experience a damaging quake. 
Right now, we get scarily little notice before the ground below starts to tremble. Earthquakes create a potent threat even in places that are well prepared — and they’ve proven catastrophic in places that aren’t. One of the deadliest disasters of the 21st century is the January 12, 2010, 7.0 magnitude earthquake in Haiti, which killed more than 220,000 people, with some estimates topping 300,000. Another is the magnitude 9.1 earthquake and ensuing tsunami on December 26, 2004, which claimed about 230,000 lives. As populations grow in seismically active areas, many more people will be at risk.
The World Meteorological Organization observed that between 1970 and 2019, the number of disasters, mostly related to weather, increased fivefold, but deaths fell by two-thirds. Even as climate change worsens downpours, floods, fires, and heat waves, improvements in weather forecasting, longer lead times for warnings, more resilient infrastructure, and better responses in the aftermath have continued to save more lives.
Yet for earthquakes, progress has been much slower. They’re a stubborn outlier in a decades-long downward trend in disaster fatalities, and recent events have shown just how deadly they can be.
But there are ways to save lives, and Japan presents an important example.