AI
Debunking AI Crowd Fakery Claims: The Truth Behind Harris’ Rally Numbers
To look over this article again, go to My Profile, and then click on View saved stories.
Claims That Kamala Harris' Rally Attendees Are Computer-Generated Are Unfounded: The Facts
In his most recent unfounded critique of Kamala Harris' run for president, Donald Trump appears to have introduced a novel phrase. Over the recent weekend, through a couple of messages on Truth Social, the ex-president accused Vice President Kamala Harris of digitally enhancing images to exaggerate the size of the audience at her campaign event at a Detroit airport the previous week, using the term "A.I.'d" to describe the manipulation.
"Trump stated, "The airplane was empty, yet she manipulated it with AI to display an enormous 'gathering' of supposed supporters, which were NON-EXISTENT! She’s engaging in DECEPTION. There was NO ONE there to greet her, and yet the 'assembly' appeared to be the size of 10,000! This pattern of fictitious 'audiences' at her rallies is recurring."
The campaign team for Harris countered by stating that the picture genuinely shows a gathering of 15,000 individuals at a Harris-Walz event in Michigan.
In a unique twist, Trump's message introduces "AI" as an action word, marking what appears to be the inaugural instance of a US presidential hopeful directly attributing the creation of artificial intelligence-driven falsifications to a rival (as opposed to the acts being attributed to political advisors or anonymous internet users). These unfounded claims exploit the prevalent anxieties and misconceptions surrounding the reliability of digital data in the era of artificial intelligence.
Ideally, we could simply dismiss Trump's statements as entirely untrue and move on. Yet, with the advancement of artificial intelligence capabilities in creating highly realistic images, it's important to detail the various methods that confirm the authenticity of Harris' crowd images. View this as a tutorial on strategies you might employ when faced with claims that an online image has been manipulated by A.I. to deceive you.
Verification and Evidence
The most straightforward method for confirming the authenticity of Harris' audience sizes is through the extensive coverage and documentation from various sources. Both the Associated Press and Getty Images have captured a multitude of images from different perspectives of the rally, in addition to coverage from journalists and those in attendance. Local media also shared footage of the gathering, with numerous participants sharing their own experiences online. Coverage from a range of media outlets provides further validation: For example, local media platform MLive pegged the crowd number at around 15,000, while The New York Times observed that the event attracted thousands, corroborating the figures provided by her campaign with what was actually observed on site.
You can also access this material on the website where it was initially posted.
It's clear that the substantial evidence provided by direct sources carries greater significance than the altered images shared by conservative figures such as Chuck Callesto and Dinesh D'Souza, both of whom have previously been implicated in disseminating false information about elections.
This article was first published on Ars Technica, a reputable platform for updates on technology, analysis of tech policies, evaluations, and beyond. Ars Technica is a subsidiary of Condé Nast, the same corporation that owns WIRED.
In dealing with claims of artificial intelligence-generated deceit, having a variety of information sources is advantageous. A lone source may produce a believable image depicting an event, yet several unrelated sources depicting the same event from different perspectives significantly reduce the chances of a coordinated fraud. Moreover, photographic evidence that corresponds with video footage is particularly valuable, given that the production of persuasive, extended videos involving humans or intricate scenarios still poses a difficulty for numerous AI technologies.
Verifying the origin of any purported AI-generated image is crucial. With the simplicity at which an individual on social media can generate an image using AI, assert it was extracted from a news broadcast or a live event, and then point out apparent defects in the doctored image to falsely claim the event was staged, vigilance is key. It's significantly more trustworthy to refer to images directly from the source, through their official website or authenticated profile, than to rely on screenshots that could have been taken from any location and possibly altered by someone.
Indicators to Watch
Searching for primary and additional sources is beneficial when reporting on significant occurrences such as a presidential gathering, but verifying the legitimacy of images and videos that come from a single source presents more of a challenge. Technologies such as the Winston AI Image Detector or IsItAI.com purport to utilize artificial intelligence models to determine the authenticity of an image. However, despite ongoing improvements in detection methods, these tools largely rely on speculative hypotheses that have yet to be validated through comprehensive research, thus carrying a tangible risk of inaccuracies.
In a LinkedIn post, UC Berkeley's Hany Farid referenced two models from GetReal Labs, stating they indicated the Harris rally images shared by Trump were not created by AI. Farid further highlighted certain areas of the photo that suggest its genuineness.
Farid notes that the inscriptions on the banners and aircraft do not display the typical markers of AI creation. He emphasizes that the absence of tampering signs does not prove the image's authenticity. Nonetheless, there's no indication that the photo has been produced by artificial intelligence or modified digitally.
Even when parts of an image seem to show illogical symbols of artificial intelligence alterations, like the oddly shaped hands often seen in AI-generated images, it's possible that there's a straightforward reason behind what looks like optical deceptions. The BBC points out that the absence of crowd reflections on the airplane in certain photos from a Harris rally might be due to a significant, vacant stretch of tarmac separating the plane from the attendees, as evident from alternative viewpoints of the location. Merely highlighting unusual features in an image with a red pen doesn't automatically serve as conclusive proof of artificial intelligence tampering on its own.
Content from X
You can also access this content on its original website.
Skepticism All Around?
As observers of AI have grown accustomed to the threat posed by AI-generated "deepfakes" and their capacity to deceive, the recent unfounded assertions by Trump about being victimized by "A.I." underscore a different facet of the deepfake issue. This aspect involves the skepticism towards genuine information due to the possibility of AI creating convincing falsehoods. This "profound skepticism" has notably surfaced among conspiracy theorists who believe that President Joe Biden has been substituted with a hologram powered by AI technology.
By prompting his supporters to question the authenticity of images that are evidently genuine (and easily verifiable), Trump seems to be aiming to foster a media landscape where the simple fact of AI's presence implies that even various credible sources of supporting evidence may not be deemed trustworthy. For a candidate who persistently makes unfounded allegations that the 2020 election was rigged against him, this could establish a perilous standard for the forthcoming election in November.
This article was first published on Ars Technica.
Suggested for You …
Direct to your email: A selection of our top stories, curated daily just for you.
A faulty update from CrowdStrike brought global computer systems to a halt
Headline Feature: The Timeline for the Atlantic Ocean's Potential Split
Introducing the age of excessive online consumption
Additional Content from WIRED
Analysis and Tutorials
© 2024 Condé Nast. All rights reserved. Purchases made via our website may result in WIRED receiving a share of the revenue, as part of our Affiliate Agreements with retail partners. Content from this site cannot be duplicated, shared, broadcasted, stored, or used in any form without explicit written consent from Condé Nast. Advertisement Options
Choose a global location
Discover more from Automobilnews News - The first AI News Portal world wide
Subscribe to get the latest posts sent to your email.