In Arizona, election workers trained with deepfakes to prepare for 2024

[wing2000]

The video message from Arizona Secretary of State Adrian Fontes warmly greeted the scores of election workers who had gathered at a Phoenix-area hotel in December for a first-of-its-kind drill: “We are very excited that all of you are here,” Fontes, a Democrat, began. “You are on the front lines, and this exercise is a prime opportunity for you to hone your skills by experiencing new challenges as a team.” He wished them luck.


Over two days, the election workers from around the state maneuvered through a training exercise involving the kinds of attacks generated by artificial intelligence that they might face in this politically competitive state during the coming election cycle. They tackled situations that plunged them into AI-generated scenarios ranging from law enforcement operations to attempts to infiltrate technology systems.
It was not until the end that the real Adrian Fontes told the shocked group that the person in the welcome video was a fake, an AI creation built out of publicly available images.

See the video here....the quality is shocking.

https://wapo.st/4dx60dr

[Some good advice at the end of the video...if a purpoted government election site does not end in .gov, don't trust it.]

 

[ThatRobGuy]

So is the primary concern then, that an AI fake would be created and used (that would look like their boss) to deliberately send them misleading instructions for tallying (via an unapproved communication channel) or something to that effect?

The AI implications I was thinking of had less to do with things that election workers have any control over, and more in the realm of misleading/fake content circulating through social media that could sway voters themselves.

 

[wing2000]

So is the primary concern then, that an AI fake would be created and used (that would look like their boss) to deliberately send them misleading instructions for tallying (via an unapproved communication channel) or something to that effect?

The AI implications I was thinking of had less to do with things that election workers have any control over, and more in the realm of misleading/fake content circulating through social media that could sway voters themselves.

I think the concern is outside actors (foreign governments, opposing political parties, etc) can create an AI fake video simulating a state's secretary of state or other election official and convey any message they want to the voters.

 

[wing2000]

...from the article:

The Arizona tabletop exercise — spearheaded by Fontes — was an effort to give one state’s election workers an insight into the types of attacks they might be up against. Following the event, organizers created a training document they plan to make available to election workers across the country, titled, “How Election Officials Can Identify, Prepare for, and Respond to AI Threats.” On Thursday, they plan to host an event in Phoenix for media to witness a similar simulation and are hoping to organize another in Michigan in the coming weeks. “But there is no way we are going to work with all 50 states, or that every election official in the country is going to be able to go through this kind of training before November,” said Norden. “That’s why we wrote the report.

 

[The IbanezerScrooge]

The AI version doesn't blink. Or only does so very sparingly. The mono-tone and slight "offness" of the speech cadence (like he's reading from a script, but even "more so") and it's just slightly out of sync with the audio (but not always) in a way that's hard to describe, but perceptible.

 

[Pommer]

The AI version doesn't blink. Or only does so very sparingly. The mono-tone and slight "offness" of the speech cadence (like he's reading from a script, but even "more so") and it's just slightly out of sync with the audio (but not always) in a way that's hard to describe, but perceptible.

Methinks a fair number of people are going to get a real world lesson in the Dunning-Kruger effect in judging these “AI deep fakes”.

 

[essentialsaltes]

The AI version doesn't blink. Or only does so very sparingly. The mono-tone and slight "offness" of the speech cadence (like he's reading from a script, but even "more so") and it's just slightly out of sync with the audio (but not always) in a way that's hard to describe, but perceptible.

The hand motion also seemed unnatural at times. But we were already aware going in. If you're just a casual or unsuspecting viewer...

I think this does a good job of emphasizing or raising awareness that AZ election workers (and everyone, really!) need to always be aware and suspicious.

 

[The IbanezerScrooge]

Methinks a fair number of people are going to get a real world lesson in the Dunning-Kruger effect in judging these “AI deep fakes”.

Probably. TBF, I watched the video knowing it was a fake so things jump out. If this were a casual watch without that knowledge it might be harder to spot. I feel like I'm decent at detecting it, but most of the time the clues come from the content and context of the video.

 

[rjs330]

The notion that only foreign actors or political organizations would do this is laughable. The government will use AI to deceive and manipulate the populace. There us great powrt in AI that is still in iys infancy stage. We should be very skeptical over what we see on video especially when it's put out by the government to sway or influence the populace.

 

[ThatRobGuy]

The notion that only foreign actors or political organizations would do this is laughable. The government will use AI to deceive and manipulate the populace. There us great powrt in AI that is still in iys infancy stage. We should be very skeptical over what we see on video especially when it's put out by the government to sway or influence the populace.

I don't think they'll necessarily use it to deceive, but the manipulation aspect could very well happen.

I think the AI-politics implication will be using deep learning to learn about what pet peeves, likes, and dislikes a person has, and then hit them with targeted ad featuring factually valid content, but in a way that draws a certain emotion or feeling.

Just as hypothetical examples....

They learn that a person is a vegan, they find a way to deliver them an authentic video clip showing "Candidate Steve Jones" eating a big old steak and saying that steak is his favorite food and how "veganism is stupid" in an off-the-cuff remark 15 years prior... followed by an authentic video clip in their feed 3 days later showing "Candidate Bill Smith" petting his dog and talking about how much he loves animals.

It wouldn't be actual deception as the clips would be genuine things that happened, but it would certainly be emotional manipulation.

 

[Handmaid for Jesus]

Methinks a fair number of people are going to get a real world lesson in the Dunning-Kruger effect in judging these “AI deep fakes”.

Yeah, like simple but honest people understand what that is.
What are the four stages of Dunning-Kruger effect?

Dunning-Kruger Effect FAQs

• Unconscious incompetence, when you don't know what you don't know.
• Conscious incompetence, when you're aware of what you don't know but haven't yet learned it.
• Conscious competence, when you're gaining knowledge.
• Unconscious competence, when you've mastered something.