<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Hamza Ali on Medium]]></title>
        <description><![CDATA[Stories by Hamza Ali on Medium]]></description>
        <link>https://medium.com/@hamzaalimian090?source=rss-0c076fb9100f------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/0*nVHd2wXPcczKjvvA</url>
            <title>Stories by Hamza Ali on Medium</title>
            <link>https://medium.com/@hamzaalimian090?source=rss-0c076fb9100f------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Fri, 20 Nov 2026 05:43:26 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@hamzaalimian090/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[How Generative Adversarial Networks (GANs) Create Realistic Images and Deepfakes and their ethical…]]></title>
            <link>https://medium.com/@hamzaalimian090/how-generative-adversarial-networks-gans-create-realistic-images-and-deepfakes-and-their-ethical-afe09c2f39f4?source=rss-0c076fb9100f------2</link>
            <guid isPermaLink="false">https://medium.com/p/afe09c2f39f4</guid>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[gans]]></category>
            <dc:creator><![CDATA[Hamza Ali]]></dc:creator>
            <pubDate>Wed, 05 Nov 2025 15:01:27 GMT</pubDate>
            <atom:updated>2025-11-05T15:01:27.422Z</atom:updated>
            <content:encoded><![CDATA[<h3><strong>How Generative Adversarial Networks (GANs) Create Realistic Images and Deepfakes and their ethical implications.</strong></h3><p>GANs have a variety of different uses across many different fields across many industries across the world. GANs are basically just an advanced deep learning model that can generate entirely new realistic data, videos, images and sound by learning from previous data sets and examples.</p><p>Why do GANs matter and how do their architectures work ?</p><p>GANs are an upcoming and integral part of AI with it being specifically in the topic of deep learning. GANs provide an entirely new way to create images and audios as a whole. This therefore provides us with newer and more efficient ways to produce images, videos etc… rather than creating them like before which was more time consuming and therefore less efficient.</p><p>GANs work by using neural network layers that gradually add detail. It does this by:</p><p>Dense (fully connected) → to build a base structure.</p><p>Upsampling / Transposed Convolutions → to make the image bigger and more detailed.</p><p>Batch Normalization → to keep training stable.</p><p>Activation functions (ReLU, Tanh) → to shape the output.</p><p>Real world examples: NVIDIA’s StyleGAN</p><p>StyleGAN stands for Style-Based Generator Architecture for Generative Adversarial Networks).</p><p>StyleGAN builds on a normal GAN generator, but further develops it by adding ‘style based” architecture. This style based architecture separates different features into high and low level features. High level features in the context of generating a human image are things like Face shape and proportions head orientation (turned, tilted, frontal), gender appearance, age, hairstyle and hair length, pose and expression, while low level features are things such as the background, skin colour the lighting and fine hair details.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/732/1*JBNNC8O0_7GAv7onivZa1Q.png" /></figure><p>GANs and their ethical impacts.</p><p>Imagine a world where anyone can create any image or video they want whether it be for good or bad, just at a click of a prompt. This is now becoming possible thanks to GANs and AI. GANs and StyleGAN have many legitimate uses in Art and other fields, but also lead to one of the most controversial technologies there are today which is deepfakes.</p><p>Deepfakes have many different ethical implications such as spreading misinformation, violating people’s privacy and making people lose trust in media sources.</p><p>Misinformation can be spread for example by producing fake videos of politicians and public figures. This can lead to false information being spread on social media which ultimately leads to loss of trust as some people don’t know what is true and what isn’t due to the more advanced technologies.</p><p>GANs also can violate many people’s privacy as certain AI tools such as DeepFaceLab and others take people’s faces for data use and allow for people to create images or videos of them without their consent.</p><p>There have also been many incidents of GAN’s being used for the worse by misrepresenting politicians to spread false news and propaganda. An example was when the New Hampshire primary elections were happening, and many people got fake phone calls from Joe Biden telling them to save their votes for November. This was a clear example of how AI can be used to interfere in politics and elections.</p><p>Attempts to regulate Deepfakes</p><p>There have been many attempts to regulate deepfakes by both companies and governments. An example of this is the EU AI act in 2024 which aimed to limit deepfakes for the better. Another example is companies such as Google, Adobe and Open AI are also trying to limit deepfakes and their privacy violations by watermarking AI generated images or videos to prevent spread of misinformation. They are also attempting to develop certain tools that will detect minor inconsistencies in GANs and deepfakes which will be able to tell what is true and what isn’t, but as GANs are developing very quickly detection of deepfakes becomes harder.</p><p>Positive uses of GANs</p><p>GANs have many different positive uses ranging from Medical usage to art and even to education.</p><p>GANs can help in medicine by generating medical images to increase datasets size and to improve image quality through super resolution which can help to train models and doctors without taking data directly from patients and other sources helping to keep their privacy.</p><p>GANs can also be used in Art . It can be used to generate unique images based on a user’s request. An example of this is a person asking AI to make a puppy walking on candy, where AI would take this input and would output a picture of a puppy walking on candy. As seen below…</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/324/1*9xYPtqd1PNJ1jG19yfoa8w.png" /></figure><p>As you can see these images made by (<a href="https://www.imagine.art/">https://www.imagine.art/</a>) are extremely realistic and are created by GANs as per the users prompt.</p><p>GANs can also be used in Education and research by simulating real world data sets and ideas for researchers when real examples are scarce and not available.</p><p>The ability of generative adversarial networks, GANs, that are currently being used, To produce convincing, fake videos and photo realistic, human faces from the very scratch, make it an important and potent form of AI.</p><p>Conclusion Pros/Cons</p><p>There are two sides of this technology:</p><p>It provides an incredible art and design tool, it makes us think critically about the truth in our digital life these days by helping artist’s, scientists and even doctors, but this could largely be misused. That’s why we need to be smart and careful and in the current world that is full of digital illusion. We need to be thoughtful and responsible more than ever, and there should be stricter rules and regulations around this.</p><p>However, how can we trust whatever we see these days online because the computer can create a flawless image of an unreal event or a video of someone speaking words they have never said?</p><p>One more important part is the consent of someone whose information and images are being used by AI. Journalism has kind of become questionable. There should be an element of accountability as someone’s reputation might be at stake or some fake videos can pause a threat to someone. For recommendations that are being shown in the social media which he never said about and may hurt someone’s feeling or maybe does not sit well with some people’s beliefs.</p><p>About me and my journey through this article.</p><p>I am Hamza Ali, I am 13 years old and I am studying in the UAE in grade 9 in an IB system school called North London Collegiate School Dubai. My journey in writing this article was through many different ways. I used my prior knowledge, researched and experimented with GAN generators myself to gain the critical knowledge that would allow me to write about GANs. I chose to write about GANs because I found them very interesting and was genuinely intrigued by them and how AI is rapidly advancing and is able to create such hyper realistic images which can be used for both good and bad.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/284/1*NC956IRGDPvhZiQtGZmDPg.png" /></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=afe09c2f39f4" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>