Perception in AI as a ‘Nice Machine’ might weaken nationwide safety disaster responses, new analysis finds

March 19, 2025

The GIST Editors' notes

This text has been reviewed in accordance with Science X's editorial course of and insurance policies. Editors have highlighted the next attributes whereas guaranteeing the content material's credibility:

fact-checked

trusted supply

proofread

Perception in AI as a 'Nice Machine' might weaken nationwide safety disaster responses, new analysis finds

Synthetic intelligence designed to affect our selections is all over the place—in Google searches, in on-line procuring solutions and in film streaming suggestions. However how does it have an effect on decision-making in moments of disaster?

Virginia Commonwealth College researcher Christopher Whyte, Ph.D., investigated how emergency administration and nationwide safety professionals responded throughout simulated AI assaults.

The outcomes, revealed within the Journal of Homeland Safety and Emergency Administration, reveal that the professionals had been extra hesitant and uncertain of their skills when confronted with utterly AI-driven threats than when confronted with threats from human hackers or hackers who had been solely assisted by AI.

"These outcomes present that AI performs a significant function in driving contributors to change into extra hesitant, extra cautious," he mentioned, "besides underneath pretty slender circumstances."

These slender circumstances are most regarding to Whyte, an affiliate professor in VCU's L. Douglas Wilder Faculty of Authorities and Public Affairs.

Nationwide safety organizations design their coaching applications to chop down on hesitancy in moments of uncertainty. Whereas a lot of the nearly 700 American and European professionals within the examine thought AI might enhance human skills, a small group believed AI might ultimately absolutely change their occupation, and human experience on the whole. That group responded recklessly to the AI-based menace, accepting dangers and rashly forging forward.

"These are those that imagine the totality of what they do—their skilled mission and the institutional mission that they help—could possibly be overtaken by AI," Whyte mentioned.

Synthetic intelligence: The following 'Nice Machine'

Whyte has a idea for why that could be the case.

The discredited "Nice Man" idea proposes that the course of historical past has primarily been formed by sturdy political figures, whereas fashionable historians give extra credit score to widespread actions. Now, Whyte proposes that historical past has additionally been formed by transformative technological innovations, just like the telegraph or radio waves, and misplaced religion of their energy—what he has coined the "Nice Machine" idea.

However not like the "Nice Man" idea, Whyte mentioned, "Nice Machines" are a shared, societal power that may be harnessed for society's profit—or for its detriment.

"Within the mid-Thirties, for example, we knew that radio waves had a large amount of potential for lots of issues," Whyte mentioned. "However one of many early concepts was for loss of life rays—you would fry your mind, and so forth."

Loss of life rays caught on, inspiring each science fiction tales and real-life makes an attempt to construct them throughout World Conflict I and the interwar interval. It wasn't till a couple of years earlier than World Conflict II that scientists started to construct one thing sensible with radio waves: radar.

Society at present faces the identical drawback with AI, Whyte mentioned, which is what he calls a "common objective" expertise that would both assist or damage society. The expertise has already dramatically modified how some individuals take into consideration the world and their place in it.

"It does so many various issues that you simply actually do have this emergent space of alternative mentalities," he mentioned. "As in, the world of tomorrow will look utterly totally different, and my place in it merely received't exist as a result of [AI] will basically change all the pieces."

That line of considering might pose issues for nationwide safety professionals as the brand new expertise upends how they consider their very own skills and adjustments how they reply to emergency conditions.

"That’s the type of psychological situation the place we sadly find yourself having to throw out the rulebook on what we all know goes to fight bias or uncertainty," Whyte mentioned.

Combating 'Skynet'-level threats

To check how AI impacts professionals' decision-making skills, Whyte recruited nearly 700 emergency administration and homeland safety professionals from the US, Germany, the UK and Slovenia to take part in a simulation recreation.

Through the experiment, the professionals had been confronted with a typical nationwide safety menace: A international authorities interfering in an election of their nation. They had been then assigned to considered one of three situations: a management situation, the place the menace solely concerned human hackers; a situation with mild, "tactical" AI involvement, the place hackers had been assisted by AI; and a situation with heavy ranges of AI involvement, the place contributors had been informed that the menace was orchestrated by a "strategic" AI program.

When confronted with a strategic AI-based menace—what Whyte calls a "Skynet"-level menace, referencing the "Terminator" film franchise—the professionals tended to doubt their coaching and had been hesitant to behave. They had been additionally extra more likely to ask for extra intelligence info in contrast with their colleagues within the different two teams, who usually responded to the state of affairs in accordance with their coaching.

In distinction, the contributors who thought of AI as a "Nice Machine" that would utterly change them acted with out restraint and made selections that contradicted their coaching.

And whereas expertise and training helped reasonable how the professionals responded to the AI-assisted assaults, they didn't have an effect on how the professionals reacted to the "Skynet"-level menace. That might pose a menace as AI assaults change into extra frequent, Whyte mentioned.

"Folks have variable views on whether or not AI is about augmentation, or whether or not it truly is one thing that's going to interchange them," Whyte mentioned. "And that meaningfully adjustments how individuals will react in a disaster."

Extra info: Christopher Whyte, Synthetic Intelligence and the "Nice Machine" Drawback: Avoiding Expertise Oversimplification in Homeland Safety and Emergency Administration, Journal of Homeland Safety and Emergency Administration (2025). DOI: 10.1515/jhsem-2024-0030

Supplied by Virginia Commonwealth College Quotation: Perception in AI as a 'Nice Machine' might weaken nationwide safety disaster responses, new analysis finds (2025, March 19) retrieved 19 March 2025 from https://techxplore.com/information/2025-03-belief-ai-great-machine-weaken.html This doc is topic to copyright. Aside from any truthful dealing for the aim of personal examine or analysis, no half could also be reproduced with out the written permission. The content material is offered for info functions solely.

Discover additional

Examine reveals white dwarfs might host life-supporting planets 5 shares