Cultura y noticias hispanas del Valle del Hudson
One of these two images was created with the help of artificial intelligence. If you thought it was the image on the left, as reported by thousands of users on social media, we invite you to keep reading this article, as it is actually the real image.
Artificial Intelligence in Elections: New Tools for Disinformation
Por Andrés Pérez Rangel
September 2024One of these two images was created with the help of artificial intelligence. If you thought it was the image on the left, as reported by thousands of users on social media, we invite you to keep reading this article, as it is actually the real image.
With the popularization of artificial intelligence tools, new waves of disinformation have emerged across all major digital platforms, which can impact serious matters such as local, state, and even presidential elections, especially in non-English-speaking communities.
Learn more about these waves and how to identify AI-generated content with this mini guide.
Artificial Intelligences are systems of algorithms designed to perform tasks such as pattern recognition, problem-solving, decision-making, language comprehension, among others. They are not limited to language: artificial intelligences are also capable of analyzing every pixel of an image, every microsecond of a video, and every sample of audio. One of their greatest features is that they can evaluate the efficiency and accuracy of each element they produce by comparing it with real elements, making AI-generated content increasingly realistic and difficult to distinguish.
The proliferation of AI-based content creation tools has increased the amount of fake images and "deepfake" videos, where a video is generally altered by imposing people's faces onto pre-existing videos; serving as a vehicle for disinformation to the general public.
According to a study conducted by the AI company Sensity, between 2019 and 2020, the flow of deepfake videos on the internet increased by 330%, with 50% of these produced or targeted at the United States. This trend is expected to continue growing. This phenomenon is showing its effects in the political sector, with recent cases of images and videos of political figures that aim to affect their public image. For example, images of Donald Trump posing with African American individuals were created to generate sympathy with that community, or a deepfake video where President Biden declared war on Russia, both completely false.
Pilar Marrero, associate editor of Ethnic Media Services, shared that "What used to require a studio budget and a production team to produce can now be assembled with a few clips in a matter of minutes for free, according to news reports. The result is that, every day, voters are trying to navigate an electoral landscape where it is difficult to assess the authenticity of images, posts, and videos."
While everyone is susceptible to exposure to deepfake content, many minority communities are more vulnerable to this kind of disinformation. Jinxia Niu, program manager for Chinese Digital Engagement, explained during the information session that most fact-checking media operate in English. Therefore, it is more challenging for non-English-speaking communities to access debunked information.
Jonathan Mehta Stein, executive director of California Common Cause, warns that the impact of this content mostly affects local politics, as organizations that debunk false content focus on high-profile political figures like Donald Trump, Joe Biden, and Kamala Harris, leaving out most local mayors and legislative members.
- Inconsistencies in texture and lighting.
- Issues with edges and blurred backgrounds.
- Excessively smooth, flawless skin and objects.
- Anomalies in text or complex designs.
- Distorted reflections or uneven shadows.
- Extra or missing fingers in hands, and mouths not synced with voices.
For the Spanish-speaking community, Factchequeado is a non-profit, non-partisan organization dedicated to countering disinformation in Spanish in the United States. They have a WhatsApp channel for inquiries and a website: factchequeado.com
back to top
COPYRIGHT 2024
La Voz, Cultura y noticias hispanas del Valle de Hudson
Learn more about these waves and how to identify AI-generated content with this mini guide.
Artificial Intelligences are systems of algorithms designed to perform tasks such as pattern recognition, problem-solving, decision-making, language comprehension, among others. They are not limited to language: artificial intelligences are also capable of analyzing every pixel of an image, every microsecond of a video, and every sample of audio. One of their greatest features is that they can evaluate the efficiency and accuracy of each element they produce by comparing it with real elements, making AI-generated content increasingly realistic and difficult to distinguish.
The proliferation of AI-based content creation tools has increased the amount of fake images and "deepfake" videos, where a video is generally altered by imposing people's faces onto pre-existing videos; serving as a vehicle for disinformation to the general public.
According to a study conducted by the AI company Sensity, between 2019 and 2020, the flow of deepfake videos on the internet increased by 330%, with 50% of these produced or targeted at the United States. This trend is expected to continue growing. This phenomenon is showing its effects in the political sector, with recent cases of images and videos of political figures that aim to affect their public image. For example, images of Donald Trump posing with African American individuals were created to generate sympathy with that community, or a deepfake video where President Biden declared war on Russia, both completely false.
Pilar Marrero, associate editor of Ethnic Media Services, shared that "What used to require a studio budget and a production team to produce can now be assembled with a few clips in a matter of minutes for free, according to news reports. The result is that, every day, voters are trying to navigate an electoral landscape where it is difficult to assess the authenticity of images, posts, and videos."
While everyone is susceptible to exposure to deepfake content, many minority communities are more vulnerable to this kind of disinformation. Jinxia Niu, program manager for Chinese Digital Engagement, explained during the information session that most fact-checking media operate in English. Therefore, it is more challenging for non-English-speaking communities to access debunked information.
Jonathan Mehta Stein, executive director of California Common Cause, warns that the impact of this content mostly affects local politics, as organizations that debunk false content focus on high-profile political figures like Donald Trump, Joe Biden, and Kamala Harris, leaving out most local mayors and legislative members.
Resources
Mehta Stein urges social media users to verify suspicious or impactful news with reliable media outlets to determine if the information is real before sharing or posting it. He also recommends the following tips to identify if an image was created by artificial intelligence:- Inconsistencies in texture and lighting.
- Issues with edges and blurred backgrounds.
- Excessively smooth, flawless skin and objects.
- Anomalies in text or complex designs.
- Distorted reflections or uneven shadows.
- Extra or missing fingers in hands, and mouths not synced with voices.
For the Spanish-speaking community, Factchequeado is a non-profit, non-partisan organization dedicated to countering disinformation in Spanish in the United States. They have a WhatsApp channel for inquiries and a website: factchequeado.com
back to top
COPYRIGHT 2024
La Voz, Cultura y noticias hispanas del Valle de Hudson
Comments | |
| Sorry, there are no comments at this time. |