<?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 Anya Macdougall on Medium]]></title>
        <description><![CDATA[Stories by Anya Macdougall on Medium]]></description>
        <link>https://medium.com/@anyamacdougall?source=rss-cf543dd139fc------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*NemAZL0YaoWWOzCWtN63QQ.png</url>
            <title>Stories by Anya Macdougall on Medium</title>
            <link>https://medium.com/@anyamacdougall?source=rss-cf543dd139fc------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sun, 22 Nov 2026 15:10:19 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@anyamacdougall/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[Measuring the Environmental Impact of AI-Generated Imagery in Creative Projects]]></title>
            <link>https://medium.com/@anyamacdougall/measuring-the-environmental-impact-of-ai-generated-imagery-in-creative-projects-ea94436beda6?source=rss-cf543dd139fc------2</link>
            <guid isPermaLink="false">https://medium.com/p/ea94436beda6</guid>
            <category><![CDATA[carbon-emissions]]></category>
            <category><![CDATA[comfy-ui]]></category>
            <category><![CDATA[generative-ai-tools]]></category>
            <category><![CDATA[creative-process]]></category>
            <dc:creator><![CDATA[Anya Macdougall]]></dc:creator>
            <pubDate>Sun, 23 Mar 2025 19:09:14 GMT</pubDate>
            <atom:updated>2025-03-23T19:09:14.461Z</atom:updated>
            <content:encoded><![CDATA[<h3>Measuring the Environmental Impact of AI-Generated Imagery in Hybrid Creative Projects</h3><p><strong>Abstract</strong><br>This article discusses the carbon emissions generated during the process of creating AI-generated orchid imagery using ComfyUI, an Open-Source image generation tool for a partially synthetic photo series titled “Digital Body Garden.” The objective of this research was to calculate the environmental impact of the generative process and propose potential offsetting methods. The findings provide an estimation of the carbon footprint associated with the open-source tool ComfyUI, presenting a mindful approach for creatives or AI artists to consider their own digital footprint when using Generative AI tools in the creation process.</p><p><strong>Keywords:</strong> <em>Artificial Intelligence</em>, <em>Open-Source</em>, <em>Carbon Footprint, Image Generation</em>.</p><p><strong>Introduction</strong><br>The never-ending development of artificial intelligence technologies has had both positive and negative impacts on the creative industry. It introduced new ways for creative brainstorming, workflow enhancements, or efficiencies, completely expanding current creative possibilities. However simultaneously, these advancements have caused many to lose their jobs and have introduced concerns regarding energy consumption and environmental impact. As the adoption of Generative AI tools increases, it’s important to investigate their ecological cost and consider solutions for reducing the emissions generated. This article focuses on the creation of AI-generated orchid imagery using ComfyUI in the creative process and the measurement of carbon emissions using an open-source tool, CodeCarbon.</p><p><strong>Methodology<br></strong><em>“Digital Body Garden”</em> is a partially synthetic photo series that combines real photographs with AI generated orchid imagery. The images were generated using ComfyUI, an open-source image generation software, a node based system also known as the Graphical User Interface of Stable Diffusion. Both software tools can be installed locally, and they are highly dependent on having an optimal graphics card, preferably with a minimum of 16 GB VRAM, utilizing the GPU for image processing.<br>The primary objective was to estimate the carbon footprint produced by the image generation process and its power usage, raise awareness about the emissions created and research ways to offset the amount generated — hopefully becoming part of best practices when working with Generative AI tools.</p><p><strong>Hardware and Software Setup<br></strong>The hardware used for the project was an NVIDIA RTX 4090 GPU. The ComfyUI software, available on GitHub, is launched by running a .bat file, which then triggers a terminal window displaying the loading process. Using the nodes and Flux Schnell as the base model, I created the workflow.</p><p><strong>Estimating the Carbon Emissions<br></strong>The ComfyUI generation process was partially monitored using CodeCarbon, an open-source Python package designed to estimate carbon footprint by measuring hardware electricity consumption. Rather than actively monitoring the entire image generation process, CodeCarbon was run for the first image generation and then tested again for a few subsequent images to compare power consumption levels.<br>The purpose of non-active monitoring was to determine whether there were significant differences in power consumption between the initial generation and the following ones. The results showed that after the first generation, the power consumption remained relatively consistent across subsequent generations. This difference was due to ComfyUI’s initial model loading and workflow initiation, which took longer compared to the subsequent generations that used pre-loaded models.<br>The calculations were based on the Netherlands’ carbon intensity rate, also called the grid emission factor, of 455 g CO₂/kWh (recorded on 26/02/2025). The monitoring method involved running CodeCarbon during the generation process and retrieving the electricity usage data from the software’s CSV output.</p><p><strong>Data Collection<br></strong>Two scenarios were analyzed using CodeCarbon and a calculator on my phone:<br><em>Initial image generation:</em></p><ul><li><strong>Generation time: </strong>Around 2 minutes.</li><li><strong>Total electricity used: </strong>0.002296 kWh</li></ul><p><em>Subsequent image generations:</em></p><ul><li><strong>Generation time: </strong>Around a few seconds per image.</li><li><strong>Total electricity used: </strong>0.001149 kWh</li></ul><p>The total number of images generated was 89, out of which 5 were selected for the final project.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*YtxkQ4qQc4VUI5Ve4yE8_g.png" /><figcaption><em>Screenshot of the Image generation workflow in ComfyUI (colorized).</em></figcaption></figure><p><strong>Carbon Footprint Calculation<br></strong>The carbon footprint calculations were made using the following formulas:</p><p><em>Emissions per image = kWh of generated image × Netherlands’ carbon intensity rate</em></p><p>0.002296 kWh × 455 g CO₂/kWh = 1.04468 g CO₂</p><p>0.001149 kWh × 455 g CO₂/kWh = 0.522795 g CO₂</p><p><em>Total emissions = emission of first image + (emission per image × number of remaining Images)</em></p><p>1.04468 + (88 × 0.522795) = 47.05064 grams of CO₂</p><p><strong>Converted to kg: </strong>0.04705 kg CO₂</p><p>To provide perspective, a standard email, without attachments, generates approximately 4 grams of CO₂. Therefore, the emissions generated during the image generation process equate to approximately sending 12 emails.</p><p><strong>Discussion and Results<br></strong>While the overall emission of 47.05 grams of CO₂ may appear insignificant, the increase in AI usage suggests that cumulative emissions from projects could become a significant environmental issue. The calculation provides a flexible method for assessing the ecological cost of AI-generated imagery. The use of local hardware and open-source software adds complexity to the measurement process, as power consumption may vary depending on hardware specifications, image generation models, and software configurations.</p><p><strong>Offsetting Measures<br></strong>One potential way to offset the emissions generated from this project is to plant a physical, “non-digital” orchid, which resonates with the project’s botanical focus. While some unverified sources suggest that orchids may absorb up to 480 grams of CO₂ annually, there is a lack of peer-reviewed research to confirm this figure. Even if that estimate were accurate, a single orchid would theoretically offset the project’s approximate 47 grams of CO₂. However, orchids are challenging to cultivate, requiring precise temperature control, high humidity, or a special kind of porous “soil.” Considering the unreliable data and the uncertainty surrounding orchids’ exact CO₂ absorption rates, planting other houseplants or donating to an orchid conservation program, such as The Mathers Foundation, maybe a more accessible option. This approach not only helps address carbon emissions in a more reliable manner but also supports ongoing efforts to preserve and study beautiful orchid species.</p><p><strong>Limitations</strong><br>CodeCarbon was not continuously run throughout the entire generation process, which may lead to minor discrepancies in total power usage estimates. Emissions also vary according to GPU model, power supply efficiency, and system configurations. Different base models and parameter configurations further influence resource consumption levels. Additionally, countries’ carbon intensity rates fluctuate, and other background processes on the hardware could have altered the overall electricity consumption. Consequently, the measurements presented here should be regarded as approximations rather than definitive values.</p><p><strong>Conclusion<br></strong>By integrating these considerations into AI-assisted creative projects, users can demonstrate a responsible approach towards AI usage. Although the estimated emissions from the image generation project may appear minimal, the cumulative effect of Generative AI usage could significantly contribute to carbon emissions. Future work should explore methods to standardize emission measurements across various hardware and software configurations, creating a reliable framework for assessing AI-generated content’s environmental impact. I hope that by sharing my insights, I can inspire AI users and creatives to become more mindful and to think more sustainably and consciously about their digital footprint.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*b5koM4GB-umw6QdMecwq8g.png" /><figcaption>Part I of the “Digital Body Garden” photo series by <a href="https://www.instagram.com/pixel.anya/"><em>Anya Victoria Macdougall</em></a><em>.</em></figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*QqqsFvU9gHTljRUfUD-llg.png" /><figcaption>Part II of the “Digital Body Garden” photo series by <a href="https://www.instagram.com/pixel.anya/"><em>Anya Victoria Macdougall</em></a><em>.</em></figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/867/1*KTPOxpyodw7B3E85cMwPrQ.png" /><figcaption>Part III of the “Digital Body Garden” photo series by <a href="https://www.instagram.com/pixel.anya/"><em>Anya Victoria Macdougall</em></a>.</figcaption></figure><p><strong>Bibliography</strong><br>Bottom, Sue. <em>Plant Growth Processes</em>. St. Augustine Orchid Society. <a href="https://www.staugorchidsociety.org/">https://www.staugorchidsociety.org</a></p><p>Orchid conservatory — donate here: <a href="https://www.justgiving.com/themathersfoundation">https://www.justgiving.com/themathersfoundation</a></p><p>ComfyUI <em>[Image generation software]</em>. <a href="https://github.com/comfyanonymous/ComfyUI">https://github.com/comfyanonymous/ComfyUI</a></p><p>CodeCarbon <em>[Carbon emission calculator].</em> <a href="https://github.com/mlco2/codecarbon">https://github.com/mlco2/codecarbon</a></p><p>Survey on Carbon Dioxide Emissions Through Email Conversion. (January 2023). SAMRIDDHI: A Journal of Physical Sciences, Engineering and Technology, 15(01), 122–127. <a href="https://doi.org/10.18090/samriddhi.v15i01.16">https://doi.org/10.18090/samriddhi.v15i01.16</a></p><p>Photosynthesis: <em>How Orchids Feed Themselves</em>. American Orchid Society. <a href="https://www.aos.org/orchids/photosynthesis-how-orchids-feed-themselves">https://www.aos.org/orchids/photosynthesis-how-orchids-feed-themselves</a></p><p><em>Effectiveness of Indoor Plant to Reduce CO₂ in Indoor Environment.</em> Mohd Mahathir Suhaimi, A.M. Leman, Azizi Afandi, Azian Hariri, Ahmad Fu’ad Idris, S.N. Mohd Dzulkifli, &amp; Paran Gani. Faculty of Engineering Technology, Faculty of Mechanical and Manufacturing Engineering, and Faculty of Civil and Environmental Engineering, Universiti Tun Hussein Onn Malaysia.</p><p>Heikkilä, M. (2023, December 1). Making an image with generative AI uses as much energy as charging your phone: <em>This is the first time the carbon emissions caused by using an AI model for different tasks have been calculated.</em> MIT Technology Review.</p><p>Heikkilä, M. (2022, November 15). <em>Why we need to do a better job of measuring AI’s carbon footprint.</em> MIT Technology Review. <a href="https://www.technologyreview.com/2022/11/15/1063202/why-we-need-to-do-a-better-job-of-measuring-ais-carbon-footprint/">https://www.technologyreview.com/2022/11/15/1063202/why-we-need-to-do-a-better-job-of-measuring-ais-carbon-footprint/</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ea94436beda6" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How the concept of existentialism and human mortality will change thanks to artificial intelligence.]]></title>
            <link>https://medium.com/@anyamacdougall/how-the-concept-of-existentialism-and-human-mortality-will-change-thanks-to-artificial-intelligence-7d29311f90a6?source=rss-cf543dd139fc------2</link>
            <guid isPermaLink="false">https://medium.com/p/7d29311f90a6</guid>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[immortality]]></category>
            <category><![CDATA[society]]></category>
            <dc:creator><![CDATA[Anya Macdougall]]></dc:creator>
            <pubDate>Fri, 01 Jul 2022 10:01:11 GMT</pubDate>
            <atom:updated>2022-07-01T11:59:29.983Z</atom:updated>
            <content:encoded><![CDATA[<p>The theory of existentialism cannot be clearly defined. In this article, I will mainly focus on the aspect of life after death, how artificial intelligence will extend human life and how it will affect the medical sector and our values. Death according to existential philosophy, pioneered by Søren Kierkegaard, is presented as the absolute end of human life. According to existentialists, death enables man to be self-aware, and the fear of it, limits him from properly “existing” on Earth. Humans are afraid of death and many are unable to accept our inevitable ‘finitude’ of the body and mind. Artificial Intelligence ‘claims’ otherwise.</p><p>Artificial Intelligence is growing rapidly, and with it, the transport industry, the marketing sector and healthcare are the most developed. With the help of machine learning, models are being developed that predict a patient’s illness based on their symptoms, allowing the disease to be diagnosed at an early stage, thereby reducing deaths and extending human life expectancy. A team of doctors researching longevity, suggest that there will be a new field of medicine called longevity medicine, which will be powered by Artificial Intelligence. Startup Insilico Medicine is on a mission to discover new medicines, which usually takes many years, wants to find treatments for many incurable diseases and in doing so support the pharmaceutical companies that are struggling. He predicts that, with the help of the previously mentioned innovative tools, people’s life expectancy will increase over the next few years. My first question is: What will life expectancy look like for long-lived people? Below, I will highlight some hypothetical situations that present different possibilities for such an outcome.</p><p>The fact is that ageing does not make us physically stronger or increase our immunity. Older age is associated with a slowing down of metabolic processes, a decline in immunity, reduced physical fitness and mobility. Just because people will have the chance to live longer does not mean that they will function and live better, in terms of quality. In addition to the practical approach to the previous situation, I must also distinguish possible situations of an emotional nature related to human longevity. The better version is, that family and loved ones will be able to spend time together and support each other for longer. The worse version would be a situation of a single, lonely person with no family. Will longer-lived people be happier? Professor of Psychology, Shimon Edelman, believes that our identity, knowing that it is partly constructed from accumulated life experiences, is an integral part of feeling happy. New experiences bring us more value in life and make us happier as people who want to experience and explore. When new experiences run out, boredom sets in. Chronic boredom can lead to a deterioration of a person’s psychological well-being and often lead to risky or self-destructive behaviour. In addition to this, society will have to adapt to demographic changes, increase the retirement age, build more retirement homes, accommodate the material and entertainment needs of the elderly, provide longer health insurance and, incidentally, have the funds for all this. Such a change will affect almost all spheres of people’s lives and will have to drastically quickly adapt to it.</p><p>It is difficult to determine whether living longer is the key to happiness and whether people need it. Working on how not to die is the most human thing about us. People are afraid of death and want to overcome it, escape it. We feel threatened by the existence of the end of our own world, which we have been creating and learning about our whole lives. If we were to be immortal, what would follow?</p><p>Cryonics is the practice of preserving life by halting the process of dying at temperatures below zero (-196°C) with the intention of restoring life in the future through technology. Scientists have already begun to work on devices that send information from the human brain to a computer. Otherwise known as ‘Digital Immortality’. This is a process whereby the entire memory and processes of the brain will be digitised and stored in a virtual brain, thereby ‘immortalising’ the mind of a biological human being. What do I think will happen after that? Above all, I think people would lose their morals. Actions and principles would have no value. Crimes would be committed and all ‘humanity’ as we know it, would die. There would be no purpose in life, nothing to strive for. Rational thinking will pass away and there will be chaos and no sense of existence. Looking at the structure of society as we know it now, would have to change completely to accommodate immortal people. Arguably, a new philosophical trend would emerge in the field of virtual immortality and new religions would be created.</p><p>The more I write about the impact of human immortality on the world the more questions arise. Is it another word for resurrection? What happens to theological mortality in the face of technological immortality? What will happen with the existing concept of death, what will death be, what will existentialism be? Will there be a new religion of immortal people, and what will happen to the assumption of ‘eternal afterlife’ and existing beliefs? What will happen to traditions, rituals and mourning?</p><p>This could be a never ending discussion. I think Artificial Intelligence may be advancing too quickly in comparison to our society’s system. In conclusion, I believe that solutions to extend the years of human life on Earth would bring more problems than solutions. Longevity would result in an unbalanced society and drastic changes would occur in every area of life. However, I would like to support the health system and help people in need, but in a way that would not be the start of another, perhaps destructive, world.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*pU2sBlLiFBPivfJh6DTSJQ.jpeg" /><figcaption>Author: Anya Macdougall</figcaption></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=7d29311f90a6" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>