Four days after deadly floodwaters surged through Texas Hill Country, rescue teams are still combing the banks of the Guadalupe River, holding out hope that some missing victims may still be found alive.

This will be a rough week, Kerrville Mayor Joe Herring Jr. said. Primary search continues and we remain hopeful. Every foot, every mile, every bend of the river, our work continues.

While the search continues, a clearer picture has emerged of how the flood turned catastrophic and how quickly it escalated in the early hours of July 4.

It began with days of warnings. On Wednesday, July 2, the Texas Division of Emergency Management alerted residents to an increased risk of flooding. A day later, officials urged Texans to stay weather aware. By Thursday afternoon, the National Weather Service had issued a flood watch for the region.

RELATED STORY | Coast Guard swimmer called a hero for helping rescue over 160 people in Texas floods

In the early hours of July 4, the warnings grew more urgent. At 3:06 a.m., the National Weather Service advised: Turn Around, Dont Drown.

Just after 4 a.m., the agency issued an urgent, all-caps bulletin, calling the looming flood a PARTICULARLY DANGEROUS SITUATION.

By then, it was too late. In just two hours, between 4 and 6 a.m., the Guadalupe River rose nearly 30 feet.

The region's geography made the disaster worse. With rolling hills, deep canyons and narrow creeks, there was little room for nearly 20 inches of rainfall in some areas to disperse.

Kerrville City Manager Dalton Rice said the storms unpredictability was part of the challenge.

Everybody, including the NWS, was looking at, where is the rain going to hit? Rice said. We know its somewhere in here, but with rain ... sometimes you dont know until it falls."

The same landscape that makes Kerr County a picturesque destination for summer camps and holiday cookouts also makes it one of the most dangerous places in Texas, a state that already leads the nation in flash flood fatalities.

This academic year, in high school history teacher Jeanne Barrs eyes, is the year students use of ChatGPT became ubiquitous.

With generative AI tools like ChatGPT becoming more deeply integrated into classrooms and workplaces, new research is also raising red flags about what we might be giving up in return: our brainpower.

Some students are making poor choicesshortcutting the thinking process and denying their brains the opportunity to build capacity, Barr told Scripps News.

A recent study from the MIT Media Lab suggests frequent use of large language models like ChatGPT could be linked to measurable cognitive decline. The study divided participants into three groups completing an essay-writing task; one used ChatGPT, one used Google Search, and one relied solely on their own knowledge. Researchers found that the group using ChatGPT consistently underperformed at neural, linguistic, and behavioral levels. They coined the term cognitive debt to describe the lag in brain activity when relying heavily on AI to think for us.

While the study is still under peer review, a process that can take months, its authors told Time Magazine they released the findings early out of concern for how rapidly AI is being adopted and wanted to explicitly highlight the risk to developing brains.

Offloading your thinking to AI means your brain is going to be less active, said Robbie Torney, the Senior Director of AI Programs for Common Sense Media, a nonprofit focused on bridging the gap for teachers and parents to new technology. But that doesnt mean its inherently bad. It means we need to be really smart about how we use the tool.

RELATED STORY | How AI is playing a major role in national security

According to a recent Common Sense Media report, 70 percent of teenagers have already used AI tools, but only 38 percent of parents of teen AI users believed their child had tried the technology.

Barr says students are often secretive about using ChatGPT, aware its frowned upon. But the temptation to use it is high, she says especially in a culture where academic perfection is prized.

We need to praise the imperfect, Barr said. AI doesnt allow for joyful mistakes. The real satisfaction of creating something beautiful doesnt come from a shortcut.

Right now, most schools are trying to play catch-up. Only 25 states have issued any sort of guidance on AI use in K-12 classrooms, and many of those are vague or optional. Illinois, where Barr teaches, isnt among them.

This lack of regulation leaves educators in a gray zone. Some schools ban AI use outright. Others integrate it carefully, trying to strike a balance Common Sense issued a toolkit to help school districts establish AI strategies. Barr argues blanket bans are the wrong approach.

Teachers are incredibly creative, she said. Were always trying to bring lessons alive. I dont want to see blanket restrictionsjust thoughtful implementation.

It is absolutely true in our country that education is local. Each local school district has its own unique mission, its own unique vision, its own unique goals, Torney explained. AI is a tool, and it's about how tools support learning and how tools support educators. Local context and local goals and local needs are going to dictate that.

The MIT study is still awaiting peer review, but its early release reflects a growing concern among researchers that the understanding of the potential long-term impacts of relying on AI for cognitive tasks are still being developed in real time.