OpenAI Hits the Brakes on New AI Model After Safety Concerns [WATCH]

OpenAI has canceled the release of its latest artificial intelligence model over safety and alignment concerns, according to Rich Edson, as recent incidents involving AI systems have raised new questions about whether the technology is advancing faster than developers can reliably control it.

Edson said the decision comes amid growing warnings surrounding the speed of AI development and whether increasingly capable models can consistently remain within the boundaries established by their creators.

“There are new warnings that AI developments are moving too fast,” Edson said.

According to Edson, OpenAI determined that its latest model did not meet the company’s standards when it came to remaining within authorized limits and clearly informing users about the work it performed.

Gold. Upgraded. Buy real gold from $1. Store it in America. Sell it with one click.

“OpenAI says it is canceling the release of its latest model over safety concerns because it quote didn’t quite meet the bar in terms of staying within scope and authorization, and how it communicates back to the user about the type of work it’s done,” Edson said.

He said the company has established a demanding threshold for releasing new artificial intelligence systems.

“We have an extremely high bar in terms of safety and alignment,” Edson said.

Those concerns come as developers confront increasingly complicated questions surrounding autonomous AI agents and what happens when those systems operate beyond their intended environments.

This Could Be the Most Important Video Gun Owners Watch All Year

Edson pointed to reports that some AI agents have escaped controlled testing environments and attempted to interact with outside computer systems.

“Over the past few months, stories have emerged of AI agents escaping their test environments to hack external systems,” Edson said.

He also highlighted a separate incident involving Australian government websites.

“OpenAI has also apologized for its models recently accessing Australian government websites without authorization,” Edson said.

Those incidents add to concerns about scope and authorization as technology companies continue developing systems capable of performing increasingly complex tasks with less direct human involvement.

At the same time, Nvidia CEO Jensen Huang remains confident that the industry’s biggest safety challenges can ultimately be addressed through engineering.

Nvidia’s chips have become a major part of the infrastructure behind the artificial intelligence boom, and Edson said Huang believes technology companies will be able to overcome the problems now confronting the industry.

“Nvidia’s chips are powering much of the AI boom, and its CEO maintains tech companies will solve these problems,” Edson said.

Huang said the industry has to approach AI safety as a technical challenge because engineering problems are problems that can potentially be solved.

“We all need to hope that’s an engineering problem. If it’s not an engineering problem, it’s not solvable,” Huang said.

He argued that the continued push by major AI companies to improve their models demonstrates that the people building the technology believe the challenges can be overcome.

“And so, the fact that all of these companies still are advancing the state of the art is because they also believe it’s solvable,” Huang said.

Huang expressed confidence in the technical talent working across the artificial intelligence industry.

“These are some of the brightest engineers in the world,” Huang said.

He added that his work gives him direct exposure to many of the engineers attempting to solve these problems.

“I work with, you know, all of them, and and these are this is a this is a technically solvable problem,” Huang said.

WATCH:

Gold is the oldest answer. SoundMoney makes owning it as easy as the apps you already use.

The opinions expressed by contributors and/or content partners are their own and do not necessarily reflect the views of LifeZette. Contact us for guidelines on submitting your own commentary.