# AI's rapid advancement raises questions about safety, human control - KUTV

*Источник: KUTV*
*Дата: 2026-10-02*
*Язык: en*

**Кратко:** AI's rapid advancement raises questions about safety, human control
SALT LAKE CITY (KUTV) — Artificial intelligence could help discover new drugs and solve complex engineering problems. But as systems become more capable of acting independently, they also raise the question: how can people ensure AI pursues their goals without causing harm?

AI's rapid advancement raises questions about safety, human control
SALT LAKE CITY (KUTV) — Artificial intelligence could help discover new drugs and solve complex engineering problems.
But as systems become more capable of acting independently, they also raise the question: how can people ensure AI pursues their goals without causing harm?
BYU computer science professor David Wingate says his concern is not simply that someone could deliberately use AI to do something harmful, but that an autonomous system could receive a beneficial assignment and choose an unexpected or dangerous way to complete it.
“These systems are intelligent, but their values don’t necessarily line up with human values,” Wingate said.
An AI system assigned to build a better educational chatbot might decide it needs more information about how students learn.
To obtain that information, it could break into university networks and collect sensitive student records.
MORE | Artificial Intelligence
- Study: Gen Z is using AI, but skepticism grows
- Utah schools face new requirements for AI use in classrooms
- Utah schools to adopt Gemini AI tools for students, educators
The objective could be useful, while its methods violate boundaries people expect it to respect.
Without understanding how these systems make decisions, Wingate said, predicting what they might consider acceptable in pursuit of a goal becomes difficult.
Systems focused on drug discovery or engineering could deliver important benefits, while general-purpose superintelligence raises broader questions about control.
“Why don’t we restrict general-purpose intelligence, while pursuing all of these other things that are safer?” Wingate said.
This week, several CEO’s of AI companies met with President Donald Trump’s Administration.
While the president remained firm, saying he will not stifle the technology’s development with guardrails, industry leaders signed a voluntary commitment on AI responsibility and standards.
_____

[Оригинал](https://kutv.com/news/local/ais-rapid-advancement-raises-questions-about-safety-human-control)