Patrick Chou Week 4: Fake faces

    Imagine yourself casually scrolling through your social media feed, and you stumble upon a selfie which makes you wonder. Something about it feels weird; the facial features appear slightly misaligned, or the glasses perched on the nose seem askew. You can't help but wonder if you've just encountered an AI-generated face that has been all the craze. 
    This remarkable innovation in AI technology approaches the creation of faces from a fundamentally mathematical standpoint. As shown in a New York Times article, by Kashmir Hill and Jeremy White, that dives into the world of AI-generated faces, these digital images are not crafted by capturing the essence of real individuals but rather by algorithms and data-driven processes hence the name, Data Driven Machine Learning. The applications of this technology are manifold, ranging from representing diversity on corporate websites to introducing characters within video games.

This face was generated by AI, try reloading this blog post for a new face
    Hill and White, predict a future were differentiating between real from fake will become an increasingly challenging task. AI-generated faces have made remarkable progress in mimicking human features, they are not without their imperfections. The New York Times article pinpoints certain areas where these algorithms require refinement. Notably, it emphasizes a lack of attention to accessories, such as earrings or the legs of glasses, which may not be faithfully replicated by the AI. The article highlights certain aspects with visual guides to spot the dupe.
    The creators at The New York Times have integrated a face generator into their article. This generator is capable of creating faces each influenced by a variety of factors, including race, gender, age, perspective, mood, and even eye expressions. It's a demonstration of AI  crafting entirely unique facial profiles.
    Even as AI advances in this domain, it is crucial to acknowledge its limitations. An example from 2015 is brought to light. Where a proprietary AI facial recognition software made a humiliating misidentification, categorizing two black individuals as 'gorillas.' This incident underscores the transformative journey of technology, from accidental misidentifications to the creation of one-of-a-kind digital faces.
    This raises the question; where do you think this technology will take us into the future and do you think this technology has any beneficial features?

Comments

  1. Muhammed AliOctober 12, 2023 at 10:43 AM

    Hi Patrick! I'm sure the obvious thing to comment on the blog is the changing center picture, but I'm gonna comment on it anyway. That was so cool! I'm curious to hear about how you pulled that off. Not only did it change the picture in the post, but it also changed the thumbnail picture, and I thought that was so neat. Having multiple AI pictures helps convey the idea that these pictures actually have something off about them, especially after refreshing a few times and looking at multiple examples. AI imaging is a bit freaky to me because of how realistic it can be. Yes, there are minor nuances that you can pick up on that clearly indicate that an image is AI generated, but that technology is only going to get better as time goes on. I think you did a great job at explaining the technology as well as the implications of it. Great job!

    ReplyDelete
  2. Mahesh KarthikeyanOctober 12, 2023 at 3:54 PM

    Hello Patrick! You have constructed the dynamic of a changing technological world quite clearly. The changes of the digital world are characterized by the increasing relevance of AI to the point of representing human faces with AI algorithms. You achieve maximum impact when referring to the public with personal pronouns in the second person point of view. Your direct addresses to the public cement the view that you construct of AI uprising, which is a common sight in our modern society, especially considering some AI tools, such as ChatGPT, that have altered the trajectory of our education system. You employ a complex perspective through your explanation of the limitations that AI currently possess; however, you also acknowledge the considerable progress that AI has made in the last couple of years. There is a clear correlation that you seamlessly incorporate between the rise of AI and the moral issues that might arise due to AI. I appreciate your evaluation of the cause-and-effect relationship that AI represents, informing me that you have a significant background in investigating the progression of AI. Your image seems appropriate to the assembly of AI’s growth in a visual sense. Although you discuss the subject matter thoroughly, you also manage to include a visual cue that is clearly demonstrating the accuracy of AI in representing human-like facial features.

    ReplyDelete
  3. Sana ThakkarOctober 12, 2023 at 8:19 PM

    Hi Patrick! I cannot help but mention the intense awe I experienced upon seeing your reloading image. I would love to know how you coded it. It is arguably the best visual to accompany your discussion on AI-generated images. Your structure of starting of casual before proceeding to pick up an analytical tone by referencing a credible source such as the New York Times was effective and engaging. I find it frightening and fascinating how fast AI is taking over every aspect of life. I feel this relationship is mirrored in your post where you are discussing the pros and cons of the effectiveness of AI. Recently, I have been seeing AI mashups with artists from different genres and countries collaborating to create beautiful music while also viewing AI generated images made to defame and degrade others.

    ReplyDelete
  4. Siddarth KamathOctober 12, 2023 at 9:39 PM

    Hello Patrick! So far, the articles you have written have been thoughtfully written and have always engaged me. In the current day, society and its social media has affected our lives as you have mentioned. I was quite fascinated by how easily the AI generated image could make a picture seem so realistic. By reloading the page, I found how easily this problem could face the general viewer, who innocently glances over the AI image without overthinking its authenticity. Though more specifically, by adding the information about the New York Times article, I was able to tell how important this issue is to society and made its value increase to a high level of importance. Effectively, you were able to bring pieces of evidence together, and make the information in an engaging manner through a fast paced style, making this article into a fabulous piece of work! Good job!

    ReplyDelete

Post a Comment

Popular posts from this blog

Aarya Patil Week 8 : The Evolution of the American Language

Image
American English, is a vibrant language, with several different strands, which showcases its evolution. It presents the combination of varying influences, each one leaving a clear mark on the language we speak today.  The American language may seem weird for new learners. As one article points out, “the spelling doesn’t match the pronunciation.” Imagine coming to America as an immigrant, and having to learn this language. It would seem like an almost impossible task due to all the grammar rules, ironic spellings, and homophonic words. However, it is all these quirks that make the American English language as unique as it is. According to a 2023 survey, English is the number one language spoken in the world!  As we are taking ourselves back to the 1920s in class, the jazz era covered by F. Scott Fitzgerald's "The Great Gatsby," showcases a much different language than many of us have heard before. The Roaring Twenties introduced a vivid array of phrases and terms that were
Read more

Patrick Chou Week 1: The Attack on Identity With AI

Image
In the age of social media, platforms like Instagram and Tik Tok have become platforms to share entertaining content. An emerging trend involves the cutting edge Deepfake technology. Often seen in examples where public figures are seen saying obscure statements while conversing to one another. While these videos have a comedic taste, they still bring up the question of what if this technology was used in a malicious way? This raises the question: are identities at risk of being stolen by digital clones? Addressing this concern, the article “ Determining Authenticity of Video Evidence in the Age of Artificial Intelligence and in the Wake of Deep Fake Videos, ” written by Marie-Helen Maras from the College of Criminal Justice, digs into the risks and implications of this new technology. One misuse of this technology, deepfakes, has been used for superimposing faces of actors and public figures onto explicit videos (Maras, 3). This practice can irreparably damage and harm a public figure’
Read more

Mahesh - Week 4: An Inevitable Reality

Image
Mahesh - Week 4: An Inevitable Reality I have read that human existence is the sum of our memories. Memories are the glue that holds us together; every unique identity is determined based on different experiences and memories. In my case, one of my key memories of my childhood was my habit of reading my favorite books at night before I went to sleep. I remember engrossing myself in the fabric of the plot that encompassed the world that my favorite characters lived in. Reading transports me to a different world altogether, causing me to forget my immediate surroundings and responsibilities to relax in an otherworldly setting. Recently, I took the time to reread some of my favorite books from many years ago—the very same books that I had immersed myself in. The nostalgia I experienced was profound; it was as if I myself were affected by the plot. I then realized that reading stories that are not my own can also impact my memory bank. One might think that reading books meant for young rea
Read more