AI misinformation Info graphic- Patrick Murtagh

Sources

-Deep Fakes are generated images or videos that are made through a machine learning technique known as deep learning. - Deep Fakes can also be used to make entertainment. - Using algorithms the technology can make extremely realistic media. - Deep fakes can mimic world leaders and create misinformation that will interfere on a global political level.

Impact 2: Deep Fakes

-Misinformation is spread most commonly on social media. - Social Media is a target for misinformation because of the large engagement that's received on social media. - Algorithms on social media are meant to personalize feeds for its users. Personalization and confirmation bias make for a large audience of misinformation.

Impact 1: Social Media Algorithms

Intro

- Misinformation or disinformation is information used to deceive or confuse populations. - Generative AI can be used to create harmful misinformation on the internet.

AI's Creation of Misinformation

Generative AI is one of mankind's greatest inventions and has many positive uses, especially to scholars and teachers. However, gen AI has made the creation of misinformation accessible and easy, causing it to spread faster than ever. Another concern from generative AI is academic integrity and creativity. Student's have become dependent on Gen AI, often plagiarizing the AI's work. AI use in younger people is decreasing creativity, since all creative decisions are deployed to the AI. While Gen AI is great and extremely useful, it comes with its own set of concerns.

Conclusion

Algorithims

AI algorithms learn patterns from online data and use it to make decisions and predictions based on those patterns, attempting to improve. The process is like teaching a machine how to learn from experience, just like how students learn from experience and practice in the classroom. https://web.p.ebscohost.com/ehost/detail/detail?vid=19&sid=2ba8d42a-90b1-46b4-8493-88f1c4640189%40redis&bdata=JnNpdGU9ZWhvc3QtbGl2ZQ%3d%3d#AN=162832382&db=afh

Generative AI is artificial intelligence that is exceptional at creative outputs that resemble human creativity. Some of the things Generative AI is great at creating are code, music, essays, summaries, and even art, this is done by the AI using patterns from data already existing on the internet. Coherent results are made from these AI's due to their data training. While generative AI is very exciting, it however creates a lot of concern as well due to ethical concerns.

Social media algorithms tailor feeds based on users’ behavior. This reinforces bias that users are likely to be shown content that aligns with their views, whether it’s accurate or not. Misinformation that fits into a user’s preconceived notions will often appear more frequently, leading them to trust and spread it without questioning its veracity.

Conformation Bias

Legistlation

The article,“Regulating AI Deep Fakes and Synthetic Media in the Political Arena", by Norden Lawarence, makes a point about legislation on AI in information such as generated media like Deep Fakes having heavier regulations. As it is important for the playing field of elections to remain clean of misinformation. While regulation on these are important the article relays to the First Amendment as a spring for tension behind legislation. Due to the first amendment protecting artistic and satirical uses of generated media. The article argues for more transparency as a requirement in manipulated media. Otherwise generative AI will become a threat to democracy due to the misinformation being spewed by things like deep fakes.

During 2019, a deep faked video of the Former Italian Priminister was created where he could be seen, saying derogatory insults, and using hate speech. This caused a stir as it was released, however it would soon be found that it was not the Prime Minister, it was an AI manipulated video. 2024 is an important year for elections, being said, there will be a ton of misinformation and deep fake incidents due to elections. Policy makers are very worried about how deep fakes will divide and inflame populations. According to “Dangers of Deep Fakes for Parliamentarians”, Inter-Parliamentary Union, in Bangladesh influential figures have been seen spreading AI generated misinformation made with cheap and accessible tools. An influencer accuses the U.S of interfering with Bangladeshi elections. The effects of Deepfakes are concerning, they will only continue to become a tool of destabilization. https://www.ipu.org/news/news-in-brief/2024-02/dangers-deepfakes-parliamentarians.

Effects Politicaly

Generative AI is very useful and even played a small role in the research and creation of my project. However it seems there are more ethical concerns about generative AI than there are pluses. As gen AI improves, more misinformation, plagiarism, and creative dependence will take place. The average IQ of humans could decrease depending on how dependent humans become on gen AI. The future is cloudy with generative AI, but let's hope the sun comes out!

General Conculsion

While yes, deep fakes can have a creative purpose and can be used to make entertainment, 99% of the time deep fakes are mentioned it is because they have mal intent. Deepfakes should be extremely concerning to the average person due to the effects this tool may pose in the future. Countries will be destabilized, people will be scammed, peoples views will be skewed, elections will be tampered due to fake endorsements and photos, ect. The point is that Deepfakes will continue to improve on quality making them a serious threat to misinformation and disinformation in the given future. Regulation hopefully will be imposed as populations gain awareness about deep fakes.

Conclusion

Social media has become a necessity in most American's lives, often using platforms like Instagram and Tik Tok for hours. With an increase of users and consumption, misinformation created by AI will only become a bigger problem. AI detectors have not been able to keep par with the amount of misinformation online. Ignorance and Hate will grow larger as minority groups and politics are often the target of misinformation, making for a concerning future on social media.

Conclusion

Deep Fakes are generated images or videos that are made through a machine learning technique known as deep learning. Using algorithms the technology can make extremely realistic media. The dangers of misinformation and disinformation. Deep learning algorithms are similar to humans in the fact that the algorithms can learn from experience. Using these to create better photos and videos. According to the article, “What The Heck is a Deep Fake," by the University of Virginia information security team, deep learning is done by neural networks with multiple "hidden layers" that make real images into realistic fakes. The networks get better by competing with another algorithm that can detect fake images. These attributes make for results that often can not be detected by the human eye or AI, and why deepfakes are used often for misinformation. More information at https://www.ipu.org/news/news-in-brief/2024-02/dangers-deepfakes-parliamentarians.

Deep Learning Algorithms

“Dangers of Deepfakes for Parliamentarians.” Inter-Parliamentary Union. (2024). https://www.ipu.org/news/news-in-brief/2024-02/dangers-deepfakes-parliamentarians. Accessed 11/6/24. Lawrence, Norden. Weiner, Daniel. “Regulating AI Deepfakes and Synthetic Media in the Political Arena.” (2023). Brennan Center For Justice. https://www.brennancenter.org/our-work/ research-reports/regulating-ai-deepfakes-and-synthetic-media-political-arena. “What The Heck is a Deep Fake.” UVA Information Security. University of Virginia. https://security.virginia.edu/deepfakes Onifade, Abdurrahman Bello. “Looking beyond the impressions of algorithms and fact-checking in fighting online misinformation: A literature review.” Education for Information. (2023). https://web.p.ebscohost.com/ehost/detail/detail?vid=19&sid=2ba8d42a-90b1-46b4-8493-88f1c4640189%40redis&bdata=JnNpdGU9ZWhvc3QtbGl2ZQ%3d%3d#AN=162832382&db=afh

Sources