By Mia Segura
@redlineproject
AI disclosure: This piece was partially assisted by AI, which was used solely to generate a basic outline and provide imagery, a summary and to create a podcast. No AI content was taken as factual information, and all research, analysis, and writing were conducted using vetted sources.
—–
Summary: AI and deepfake technology have reshaped political journalism, introducing misinformation, ethical dilemmas and public skepticism. Deepfakes manipulate media, making it difficult for journalists to discern truth from fabrication. AI’s evolution since the 1950s has further complicated news credibility. The spread of political deepfakes influences voter opinions and fuels distrust. Legislative measures like the DEEP FAKES Accountability Act aim to regulate AI misuse. Media literacy and AI detection tools remain crucial in combating misinformation.
___
Introduction
Today, AI and deepfake technology have caused a major shift in how political news is reported in journalism. As these technologies evolve, journalists have had to learn how to navigate this new system filled with public skepticism, the spread of misinformation, and ethical concerns. Although AI was not created to be a harmful device, its ability to manipulate narratives poses a great threat to politics. This piece will explore topics of AI and deepfake technology, how they have impacted political journalism, and what journalists are doing to remain credible sources.
Understanding AI and Deepfakes
The manipulation of images has occurred for over a century, and although it has always been a problem, the creation of deepfakes has taken the issue to a whole new level. A report by the United Kingdom Government in 2019 dates the creation of deepfakes back to 2017 when a user on Reddit started posting pornographic images that were altered to look like celebrities rather than the real faces of adult entertainers.
As for the development of AI, its creation dates back to the 1950s when Alan Turing created the Turing Test. This test would evaluate the capability of a machine to show intelligent behavior that is similar to humans. As time went on, so did the research and development of artificial intelligence, and it wasn’t until the 2010s that AI began getting recognition from the public after a human was defeated by AI in the well-known game show “Jeopardy!”. From there, AI continued to grow in popularity and was used regularly.
People were quick to spread AI and deepfake content, and it soon became normalized. This normalization came with its own set of issues, one of the biggest ones being political deepfakes. The main job of journalists is to tell the truth and report the facts; however, now that any piece of media can be altered and fabricated, journalists have struggled to decide what is real and what is fake.
Misinformation and Public Trust
The presence of both AI and deepfake technology in politics is highly concerning because of their ability to shape public opinion. When voters are exposed to content that appears to be realistic of politicians discussing controversial topics, they are going to be quick to form an opinion on said politician. Their initial response is not going to question the authenticity of the media, but instead, it is going to affect their voting decisions. It is circumstances like these that create distrust and division in news sources.
Professor Gregor Baszak at the University of Illinois-Chicago gives his perspective on AI in journalism and whether he believes the general public is informed enough of its risks: “People fall for all sorts of BS and propaganda if passions run high enough and their tribe gets enraged or animated.”
With his background in publishing political writing pieces, he has had firsthand experience with being surrounded by disinformation and navigating through it.
“The best way not to fall for lies is to grow a spine, suspend judgment, listen to your inner skeptic, and tell the majority they’re wrong if you’re certain of that,” Baszak said. “Unfortunately, trust in the news media is at an all-time low, and that’s well-deserved.”
With the large threat of misinformation, the public will continue to not trust news sources, and without that trust, where else are people going to get factual information? This issue threatens the entire existence of journalism.
Implications for Political Journalism
The past few years in politics have been a rollercoaster of a journey for most voters in the United States. For all sides of the political spectrum, accusations have been made, information has been spread, and a division amongst Americans has been rapidly growing.
In January 2023, voters received a robo-call in New Hampshire that appeared to be former President Joe Biden encouraging voters not to vote in the 2024 primary election. Some believed it and stayed home. It became a national news story that serves as an example of the continuing issues surrounding generative AI.
—–
PODCAST: Listen to an AI-generated podcast about AI and deepfakes in media covered in this paper.
___
As for deepfake technology, there was an incident reported just before the 2020 election in which House Speaker Nancy Pelosi was targeted after a deepfake recording was created of her appearing to be intoxicated while giving a political speech. Being a woman in politics comes with its own set of challenges, and if AI-generated content is being released, it does even more damage to their credibility.
There is a repeated pattern when it comes to AI and deepfake usage in politics; upcoming presidential elections are occurring. The spread of altered images, information, and data happens to line up when political discussions are at their peak, thus making it impossible for consumers to gather truthful information that would help them make their final decision.
Legal Considerations
The increasing numbers of AI and deepfake usage in politics have already led to bills being introduced to Congress. Last year, 27 states were reported to have introduced said bills, and many have demanded transparency in AI usage and a ban on deepfake content released a certain number of days before elections.
Other countries have started their own legal considerations, and some have been successful. China’s Personal Information Protection Law (PIPL), which took effect in 2023, required explicit consent from those whose image, voice, or any personal data was used in synthetic data. These measures prove just how effective legislation can target and eliminate legal and ethical concerns surrounding AI and deepfake technology in the media.
As of 2023, the U.S. has introduced the DEEP FAKES Accountability Act, which is very similar to the PIPL in China; however, this act would specifically target generative AI in the political atmosphere. Prosecutors and regulators would have access to AI detection software, and anybody who creates a deepfake would be required to identify the usage. Anyone who would not follow this act could face a criminal penalty.
Education and Media Literacy
The next step to develop education on AI usage, making it a mandatory part of K-12 school curriculums. The younger generation is the first to grow up in a world full of artificial intelligence, and one of the easiest ways to make sure they learn how to navigate it would be to start educating them as early as possible. If it was a requirement, it could decrease the likelihood of someone believing fake content.
As for journalists, the best solution to ensure the spread of misinformation is avoided would be to closely follow the steps of media literacy. An article written by the NBCU Academy takes note of an acronym, “CARES,” to specifically help journalists remember steps.
The first step is to gather context. Journalists must ensure that proper information is being presented about the topic.
Next, journalists would be asked to verify authenticity. Verification of unedited information written by educated individuals is identified.
The third step would be searching for reasoning. There has to be a logical argument being made along with the other information presented.
At this point, journalists should ensure there is evidence being cited that is trusted. There must be proof to back up the argument being made.
Lastly, journalists are asked to research the sources being used and make sure they are credible.
AI Detection and Combating Deepfakes
While U.S. legislation is still working on legal actions, there are things people can do now to protect themselves. Although often referred to together, AI can be used to combat deepfakes. Media organizations and tech companies have begun developing systems that can recognize manipulated content by using artificial intelligence.
A popular example is Sentinel AI, which is a deepfake detection tool that can easily identify any visual discrepancies and altered audio. By using a program like this one, anybody can determine whether something is real or fake.
There are also tools as accessible as Microsoft’s Video Authenticator, which verifies the authenticity of video or image content by using machine learning.
The Balance
While the creation of AI and deepfakes has raised concerns about misinformation and skepticism, they also offer advantages in news reporting. AI can automate data analysis, summarize information, and even help journalists investigate and issue. There are tools powered by AI that can efficiently search through public records and uncover any sort of patterns or corruption. Rather than completely weakening journalism, some potential enhancements can be made because of AI’s efficiency.
Journalists need to improve their ability to identify misinformation while also incorporating AI tools to verify sources and maintain credibility. Fact-checking systems can detect inconsistencies and help stop the spread of false news. However, human involvement in reporting remains crucial in ensuring ethical reporting.
As for deepfakes, there is also a balance that can be incorporated. Many uses of deepfake technology have reportedly been to sabotage political opinions or ruin the public image of important people, but deepfake technology can also be useful for personal reasons.
Deepfake technology can help recreate moments from the past and make people connect with history on a deeper level. There was recently a Scottish company, CereProc, that recreated audio from a speech by former U.S. President John F. Kennedy that he was meant to give in Dallas the day of his assassination. This usage of deepfake technology is a gift to those who were not around during his presidency and also allows for people to reconnect with the president they lost too soon.
Conclusion
AI and deepfake technology have undeniably transformed political journalism, bringing in better opportunities for journalists and also bringing in difficult challenges. The spread of manipulated media has continued, making it difficult to recognize what is fact and what is fake, and this ultimately leads to the public opinion being influenced and voter decisions being determined for the wrong reasons.
Public skepticism toward media outlets will continue to grow as AI content develops, and the trust in journalism emphasizes how crucial it is for legislation to become more involved. There is a call for lawmaking, media literacy education, and an advancement of AI detection that needs to be answered. Legislation such as the DEEP FAKES Accountability Act in the U.S. can be beneficial, but it is also important to realize that this is not the only solution necessary.
Ultimately, it is the responsibility of media organizations, journalists, and even individual people to remain vigilant and analyze the content they are sharing. These types of technologies are bound to evolve, and this only highlights the need for adaptation to occur.