Can AI influence election outcomes?

Published

Artificial intelligence (AI) may be a weapon of mass disinformation, but a recent report has demonstrated that its impact thus far has been limited.

Voters in nearly 100 countries – including Taiwan, the US and Senegal – went to the polls this year, and AI was often used during the election campaigns. This technology, when used in malevolent ways such as through deepfakes and chatbots, erodes citizens’ trust in the information provided by news outlets, whether on TV, online or in social media. AI-driven programs have clearly affected the reliability of the information we receive, but has that had an impact on election outcomes? A team of researchers at the EPFL-based Initiative for Media Innovation (IMI) conducted a study to investigate the influence that AI had on elections around the world in 2024. The findings appear in the first issue of IMI’s Décryptage magazine (in French only). The issue was written by Swiss journalist Gilles Labarthe in association with Mounir Krichane and Julie Schüpbach, both at IMI, and Christophe Shenk, chair of IMI’s Scientific Committee and head of digital news coordination at Swiss broadcasting company RTS.

Resurrected political figures

The researchers worked with local experts to analyze the various election campaigns and results. They found that AI-driven programs had only a marginal impact and didn’t swing the elections one way or the other. However, the study did find that the spread of manipulated content, boosted by algorithms, divided political opinion further and created a widespread climate of mistrust. For example, deepfakes – videos that have been digitally altered so that they appear to display actual people – were used in election campaigns in both the US and Switzerland. Meanwhile, generative AI was taken to a whole new level in India and Indonesia, where programmers brought political figures back from the dead by creating avatars intended to sway voters.

“Technology on its own won’t be enough,” he says. “Human users are the weak link.”

– Touradj Ebrahimi, Head of EPFL’s Multimedia Signal Processing Group

The authors of the study stress that the use of digitally manipulated content for propaganda purposes is nothing new; AI has only amplified this practice. The large-scale production and rapid dissemination of fake content – whether in video, image or text format – during election campaigns have undermined citizens’ trust. The authors also point to a regulatory vacuum that has enabled such content to circulate freely.

In an interview for the magazine, Prof. Touradj Ebrahimi, head of EPFL’s Multimedia Signal Processing Group, says that deepfakes are creating unprecedented technical, societal and ethical challenges. “It’s a game of cat and mouse between the creators of AI technology to generate deepfakes and the developers of software to detect them.” His research group is working to develop systems for identifying and limiting the dissemination of manipulated content (see below).

A collective effort

The IMI magazine provides a sweeping view of the risks that AI poses for election campaigns. It also gives concrete recommendations from scientists, other experts and media professionals for reducing the impact of disinformation, and suggests actions citizens can take. One recommendation is to implement fake-content detection and tracing systems, like the ones being developed by Ebrahimi’s group.

The magazine highlights the importance of introducing international regulations and of holding the media accountable. For his part, Ebrahimi says it will be essential to encourage collaborative fact-checking and promote education as a powerful ally in the fight against disinformation. “Technology on its own won’t be enough,” he says. “Human users are the weak link – we’ve got to make them aware of the risks associated with fake news and give them resources for verifying the sources of the information they receive.”

Finally, the magazine underscores the crucial role that governments, businesses and civil society can play in making the digital space both ethical and secure. This will require a collective effort to restore trust in the democratic process as AI becomes ever more prevalent.

Winning the fight against disinformation will require not just developing the right technology, but also – as the IMI magazine explains – a concerted effort among scientists and engineers, governments, businesses and citizens. Together, we can make information reliable again and restore trust in the democratic process.

EPFL’s unique expertise in combating manipulated content
At Prof. Ebrahimi’s Multimedia Signal Processing Group, engineers are working to develop technology that can effectively detect and stem the spread of manipulated content. This includes implementing the JPEG Trust standard so that the authenticity of images can be verified from the time they’re created until they’re published.
“There’s no magic bullet,” says Ebrahimi. “Instead, we’ll need to combine several indicators in order to build trust and reduce the risks.” This proactive approach could entail adding digital signatures to content, for example, so that users can trace it and detect any unauthorized changes.
Ebrahimi’s group is also examining the use of generative adversarial networks (GANs), which are networks where two machine learning programs compete against each other – one produces fake content while the other aims to detect it. GANs can enhance the ability of detection technology to spot even the most sophisticated deepfakes, providing a valuable tool for online media outlets and other content platforms.

Encouraging innovation in digital media
IMI was founded in 2018 by public- and private-sector organizations to promote digital innovation in the media. Its members include EPFL, SRG SSR, Ringier and the Universities of Geneva, Lausanne and Neuchâtel. The initiative has the support of the Swiss Federal Office of Communications.

Author: Mélissa Anchisi

Source: EPFL

Share

You might be also interested in

EPFL researchers create an AI model that thinks like we do

An EPFL team has created a new Large Language Model that is structured similarly to a human brain, allowing users more control and moving away from “black box” AI.

When a standard Large Language Model (LLM) is confronted with a problem, it tries to solve it by matching it to similar information it has seen before, and then give an answer based on those past patterns. But how it decides which information to use and what value it gives to different pieces of information can be somewhat inscrutable from the outside.

The LLM MiCRo (Mixture of Cognitive Reasoners) is architecturally divided into four specialized areas that act like different parts of the human brain, allowing users to have more control over how it approaches a question, and to better understand how it comes to its answers. The model, which was presented at the International Conference on Learning Representations, comes from the NLP Lab, part of the School of Computer and Communication Sciences (IC), and the NeuroAI Lab, part of IC and the School of Life Sciences at EPFL.

The four experts

To create MiCRo, researchers identified four regions of the brains specializing in different functions, which they call ‘experts’: language, logic, social reasoning, and world knowledge.

“The brain is organized into specialized regions, each tuned to handle a specific function. So far, we don’t see this division of labor as clearly in current language models,” says Badr AlKhamissi, a PhD candidate leading this research. “We picked four brain regions that neuroscientists know well and gave the model its own specialized modules, each one trained to be analogous to one of those brain regions.”

An LLM usually functions as a stack of layers that a problem or question can be processed through. In the case of MiCRo, each layer is divided into the four different experts. You give a sentence to the model starting at layer one, for example “The cat is asleep”. Then within this layer, the router can choose one expert for the first word “the”, but a different epxert for second word “cat” and so on, making it modular and highly adaptable.

“Each word of a sentence can go to different experts,” AlKhamissi explains. “So one sentence can actually be processed by multiple experts at each layer.”

Consider a prompt like: “Emma wants to split a CHF 60 dinner bill among three friends, but she knows that Jake lost his job last week and is too proud to say he’s struggling.” A purely mathematical module handles the arithmetic: CHF 60 divided by three is CHF 20 each. But the social reasoning module picks up on something subtler: Emma’s awareness of Jake’s situation, his unspoken pride, and the implicit suggestion that she might quietly cover his share. Both kinds of reasoning are needed to fully understand what’s going on, and in MiCRo, each aspect of the prompt is routed to the expert best equipped to handle it.

“When we see how the model works, we can see that it routes the words that relate to the social aspects to the social expert, and when it does the mathematical part, it routes those numbers to the logic expert.”

This separation makes it easier to see how the model is ‘thinking’ and why it makes certain decisions. It also means decisions can be steered – for example, you can decide to increase the impact of the social expert, or suppress the logic expert, depending on what kind of model you want to use in a certain situation.

“In traditional LLMs, you can do this via prompting by telling the model to make the output more social or make it more related to emotions,” AlKhamissi says. “But here, this is done by intervening in the architecture itself without doing any prompting.”

“A virtuous circle”

To create MiCRo, the EPFL team worked with Greta Tuckute, a neuroscientist from Harvard and MIT, to understand which parts of the human brain are activated by different problems, and then applied that learning to the model.

To identify the region analogous to the ‘logic’ expert in the brain, neuroscientists give humans demanding tasks, such as hard mathematical equations, and less demanding tasks, like easy mathematical equations, and then recorded their brain activity to find which brain regions are the most active for the demanding tasks versus non-demanding tasks. AlKhamissi’s team then did the same for the model, giving it demanding mathematical equations to see which experts would be most activated.

“The cool thing is we just used exactly what they do in neuroscience, but in the model. And the model was able to identify those experts on its own.”

While neuroscience informs the model, the model also informs the understanding of the brain, potentially allowing neuroscientists to discover the contributions of different areas for a given problem or question; for example that a certain sentence activates the language areas 20%, the mathematical areas 50%, and the social reasoning areas 40%.

“For my PhD work, I have been interested in this virtuous circle between neuroscience and AI. In one direction, we use findings and insights from neuroscience about the brain and integrate them into language models,” AlKhamissi says, “and now, with models like MiCRo, we can explore the other direction and ask how we can use AI models to help us understand the brain in a better way.”

Author: Stephanie Parker
Source: EPFL

Read more

Smarter waste sorting with AI

EPFL startup WasteFlow has developed an AI-powered copilot that identifies and measures waste streams, helping sorting facilities work more efficiently. Support from several EPFL entrepreneurship programs helped the company accelerate the development of its technology.

(more…)
Read more

EPFL launches the world’s first fully open medical LLMs

MeditronFO is the first fully open framework for building medical large language models, to make AI in healthcare more transparent and accountable.

(more…)
Read more