Ethics and use of AI

AI and Its Impact on Climate, Economy, and Society

AI has revolutionized industries, especially in boosting creativity, productivity, and efficiency. In the creative sector, AI complements human innovation rather than replacing it (1) (2). While the debate about AI's role in creativity persists, many believe true creativity is human, with AI enhancing it. For more on this, check out my blog on AI & Creativity. However, AI's ethical concerns, especially its environmental and societal impact, raise significant questions.

Environmental Impact of AI

AI tools like ChatGPT raise concerns about energy use and environmental impact. Data centers consume significant power and land, creating an ethical dilemma of balancing innovation with sustainability. While boosting efficiency, they also strain power grids and increase emissions. AI systems use 33 times more energy than specific software, with GPT-3 training consuming 1,300 MWh, equal to 130 U.S. homes' annual power. GPT-4 used up to 50 times more energy for its creation. AI's rapid growth doubles power needs every 100 days, straining grids and requiring major energy investments.

To create ChatGPT-4 it used 65.000 MWh of electricity, equal to 6500 homes anually, which equals to 0.000325% of the global electricty used in 2023 (6). In my eyes not really a big deal energywise for ChatGPT specificly.

Efforts to improve energy efficiency include innovations like specialized hardware, advanced cooling, and optimized data centers. For example, Nvidia’s new ‘superchip’ boosts performance 30 times while using 25 times less energy (3):

Let’s make this story more visual to show how big of an improvement this is!

Imagine you're going on a 1,000 km vacation road trip. Normally, you drive at a speed of 100 km per hour, and for every 100 km, you use 10 liters of fuel. So, for your whole trip, you'd use 100 liters of fuel to reach your destination.

Now, let’s say you get a magic boost that makes your car 30 times faster while using 25 times less fuel. With this boost, your car can travel 30 times faster, so instead of driving 100 km per hour, it goes at 3,000 km per hour. That means you would reach your 1,000 km destination in just 20 minutes instead of 10 hours!

And here's the best part: since the car uses 25 times less fuel, it only needs 4 liters of fuel for the whole trip instead of 100 liters. So, you get to your vacation spot in a flash, using just a tiny amount of fuel, talk about an energy-efficient road trip! Conclusion, old car vs new car = (current chip vs new chip):

The old car traveled 10 hours using 100 units of fuel, while the new car traveled the same distance in 0.33 hours using only 4 units of fuel.

To create even more perspective on the energy usecase: one person in Qatar consumes 226,848 kWh of energy anually (8), which is equivalent to 0.348% of the energy used in the creation of ChatGPT-4. This means that the energy consumption of 288 people in Qatar is the same as the energy used to create ChatGPT-4.

So Yes! AI requires a lot of energy to create, but the energy used to develop ChatGPT-4 is only 0.01% of the energy consumed by dishwashers globally (9). So, should you start washing your dishes by hand instead? I hear you asking, "But isn’t this just about the creation? What about the ongoing use case worldwide?" Let’s dive into that and explore it further! Let's make a calculation for the ongoing usecase: (10) Google search uses 0.0003kwh (10) ChatGPT search uses 0.015kwh (avarage)

According to a source (11), ChatGPT processes 156.4 billion prompts annually and growing, consuming around 2.35 billion kWh, roughly 0.01175% of global energy consumption (9). For ChatGPT-4, the total energy use, including both creation (0.0003%) and ongoing usage, is just 0.01%. Despite this energy consumption, AI holds the potential to reduce global emissions by 5-10% by 2030 (4) (5) while driving advancements in climate science.

Employment and Economic Disruption (12) (13) (14)

AI threatens to replace up to 40% of jobs, especially in routine roles. While it creates new tech-related job opportunities, these often require reskilling. Without proper policies to manage the transition, economic inequality and unemployment could rise, further exacerbating social disparities.

Ethical Implications of AI-Generated Content (15) (16) (17) (18) (19) (20)

The rise of AI in the creative world sparks some intriguing questions about intellectual property, ownership, and the authenticity of AI-generated art. If AI borrows from human-created works, can artists still protect their rights? With AI’s ability to mimic artistic styles, the line between human and machine-made art is getting blurry, challenging our ideas of creativity.

When it comes to copyright, things get tricky. Traditional law protects works of human creativity, but who owns the content created by AI: the user, the developer, or no one at all? Fair use, which lets you use copyrighted material in certain situations, adds another layer of complexity. AI’s training on copyrighted works or creating inspired content often falls into a gray area of what’s considered fair.

Weaponization of AI (21) (22) (23) (24)

AI's potential use in military applications raises serious ethical concerns, particularly regarding autonomous weapons systems and the risk of escalating conflicts without human oversight. The ethical dilemmas surrounding life-and-death decision-making by machines require urgent attention, along with international regulations governing military AI use. The "Dead Internet Theory," Indirect Warfare, AI Bias (25) (26) (27) (28) (29)

The Dead Internet Theory suggests that much of online content, especially on social media, is generated by bots or AI, creating an illusion of real human interaction. This links to indirect mental warfare, subtly shaping opinions and steering society without confrontation. AI bias worsens the issue, reflecting societal prejudices and deepening inequality. As AI grows, we must ask: Are we connecting with real people or consuming artificial simulations?

The risks are also growing with AI, drones, and facial recognition. Flawed data could lead to drones targeting people, and system errors could cause catastrophic consequences. AI in warfare is terrifying, and though I hope we never see autonomous killing, the path to that future feels increasingly likely.

Conclusion on AI’s Impact on Climate, Economy, and Society

AI is revolutionizing creativity, productivity, and efficiency, but its rapid growth brings up ethical concerns, especially regarding its environmental and societal impact. While AI can enhance human creativity, we must consider its effect on the planet and society.

Environmental Impact: AI, like ChatGPT-4, uses a tiny fraction of global energy, around 0.0003% for its creation, and 0.01% for maintenance. Though minimal, AI’s growing computational demands stress power grids and contribute to emissions, highlighting the need for energy-efficient systems.

AI and Climate Action: Despite its energy use, AI can help combat climate change by optimizing energy grids, advancing climate models, and cutting emissions. But to truly help, AI needs to be energy-efficient and minimize data center emissions, with projections showing a need for 90 TWh by 2026 compared to 20,000 TWh globally.

Economic Disruption: AI threatens to replace up to 40% of jobs, especially in routine roles. While new tech jobs will emerge, they’ll need reskilling. Without proper policies, economic inequality could rise, just look at all the technological dystopian movies!

Ethical Dilemmas: In the creative world, AI raises concerns over copyright, ownership, and authenticity. We’ll need new laws, like the EU AI Act, to protect artists and ensure fairness.

Military Concerns: AI’s potential military use, especially in autonomous weapons, presents grave ethical issues. Without human oversight, we need international regulations to prevent misuse.

AI holds incredible potential, but its environmental and societal impacts need thoughtful management. With advancements in energy-efficient chips and a surge in renewable energy, the real challenge isn’t energy consumption, it’s ensuring innovation is guided by strong ethical frameworks. In just a few years, we’ll see breakthroughs that make current concerns about energy use seem small, as technology and sustainability evolve hand in hand (30) (31) (32). The real question is: will we accept the ethical costs of rapid progress, especially in areas like copyright, as we advance? Is the global benefit of faster creative ideation truly a fair use case?

If you found this interesting or have any questions,

don’t hesitate to reach out! I’d love to help, contact me today!

Sources used to create the post:

(1) Creative Machines: The Rise of Generative AI and its Impact on Human Creativity. (2023, februari). Scu. Geraadpleegd op 25 november 2024, van https://www.taylorhawkes.com/knowledge/the-ai-revolution-ethical-implications-for-the-creative-industries (2) Santa Clara University. (2024, augustus). The AI Revolution: Ethical Implications For The Creative Industries. Taylorhawkes. Geraadpleegd op 25 november 2024, van https://www.scu.edu/ethics-spotlight/generative-ai-ethics/creative-machines-the-rise-of-generative-ai-and-its-impact-on-human-creativity/ (3) Hsu, J. (2024, 20 maart). Nvidia's Blackwell AI 'superchip' is the most powerful yet. New Scientist. https://www.newscientist.com/article/2422928-nvidias-blackwell-ai-superchip-is-the-most-powerful-yet/

(4) AI and energy: Will AI help reduce emissions or increase demand? (2024, 10 september). World Economic Forum. Geraadpleegd op 25 november 2024, van https://www.weforum.org/stories/2024/07/generative-ai-energy-emissions/

(5) How AI Can Speed Climate Action. (2023, november). Bcg. Geraadpleegd op 25 november 2024, van https://www.bcg.com/publications/2023/how-ai-can-speedup-climate-action

(6) Topic: Global energy consumption. (2024, 25 januari). Statista. https://www.statista.com/topics/4042/global-energy-consumption/#topicOverview (7) Kim, J. (2023, 27 november). Four ways AI is making the power grid faster and more resilient. MIT Technology Review. https://www.technologyreview.com/2023/11/22/1083792/ai-power-grid-improvement/#:~:text=Four%20ways%20AI%20is%20making%20the%20power%20grid,4%204.%20Spotting%20disasters%20before%20they%20hit%20

(8) Energy use per person. (2023, 1 augustus). Our World in Data. https://ourworldindata.org/grapher/per-capita-energy-use?tab=table

(9) How Energy Efficient are Modern Dishwashers? (2008). fsec.ucf. Geraadpleegd op 25 november 2024, van https://www.fsec.ucf.edu/en/publications/pdf/fsec-cr-1772-08.pdf

(10) Chat, F. (2024, 25 november). ChatGPT vs. Google Search: Energy Consumption Comparison. Fusion Chat. https://fusionchat.ai/news/chatgpt-vs-google-search-energy-consumption-comparison

(11) Hoffman, P., & Hoffman, P. (2024, 15 november). AI’s Power Demand: Calculating ChatGPT’s electricity consumption for handling over 156.4 billion user queries every year - BestBrokers.com. BestBrokers.com - Best Brokers for Trading Online. https://www.bestbrokers.com/forex-brokers/ais-power-demand-calculating-chatgpts-electricity-consumption-for-handling-over-78-billion-user-queries-every-year/

(12) Liang, B. A. (2024, 15 januari). AI to hit 40% of jobs and worsen inequality, IMF says. https://www.bbc.com/news/business-67977967 (13) These are the jobs most likely to be lost – and created – because of AI. (2024, 10 september). World Economic Forum. https://www.weforum.org/stories/2023/05/jobs-lost-created-ai-gpt/ (14) Haan, K. (2024, 16 oktober). 22 Top AI Statistics and Trends in 2024. Forbes Advisor. https://www.forbes.com/advisor/business/ai-statistics/ (15) Deng, Y. (2024, 10 juli). Artists’ Rights in the Age of Generative AI | GJIA. Georgetown Journal Of International Affairs. https://gjia.georgetown.edu/2024/07/10/innovation-and-artists-rights-in-the-age-of-generative-ai/ (16) ScoreDetect Blog | Intellectual Property & Copyright Protection. (2024, 7 september). ScoreDetect Blog | Intellectual Property & Copyright Protection. https://www.scoredetect.com/blog/posts/can-you-copyright-ai-art-legal-insights

(17) news.iu.edu. (2023, 11 juli). Ask the Expert: What are legal issues surrounding AI, its impact on the arts? news.iu.edu. https://news.iu.edu/live/news/31782-ask-the-expert-what-are-legal-issues-surrounding-ai-it

(18) Artificial intelligence and copyright: use of generative AI tools to develop new content. (2024, 16 juli). IP Helpdesk. https://intellectual-property-helpdesk.ec.europa.eu/news-events/news/artificial-intelligence-and-copyright-use-generative-ai-tools-develop-new-content-2024-07-16-0_en

(19) Artificial Intelligence Act. (2024). europa.eu. Geraadpleegd op 25 november 2024, van https://www.europarl.europa.eu/doceo/document/TA-9-2024-0138_EN.pdf (20) Generative AI and Copyright Law. (z.d.). https://eucrim.eu/news/generative-ai-and-copyright-law/ (21) Militarization of AI Has Severe Implications for Global Security and Warfare. (2023, 14 augustus). United Nations University. https://unu.edu/article/militarization-ai-has-severe-implications-global-security-and-warfare (22) Dresp-Langley, B. (2023). The weaponization of artificial intelligence: What the public needs to be aware of. Frontiers in Artificial Intelligence, 6. https://doi.org/10.3389/frai.2023.1154184 (23) Understanding the Strategic Implications of the Weaponization of Artificial Intelligence. (2019, 1 mei). IEEE Conference Publication | IEEE Xplore. https://ieeexplore.ieee.org/document/8756866

(24) Deng, Y. (2024b, juli 14). AI & The Future of Conflict | GJIA. Georgetown Journal Of International Affairs. https://gjia.georgetown.edu/2024/07/12/war-artificial-intelligence-and-the-future-of-conflict/

(25) The ‘dead internet theory’ makes eerie claims about an AI-run web. The truth is more sinister. (z.d.). UNSW Sites. https://www.unsw.edu.au/newsroom/news/2024/05/-the-dead-internet-theory-makes-eerie-claims-about-an-ai-run-web-the-truth-is-more-sinister

(26) Team, I. D. A. A. (2024, 25 november). Shedding light on AI bias with real world examples. Shedding light on AI bias with real world examples. Geraadpleegd op 26 november 2024, van https://www.ibm.com/think/topics/shedding-light-on-ai-bias-with-real-world-examples (27) AI Bias - What Is It and How to Avoid It? (z.d.). https://levity.ai/blog/ai-bias-how-to-avoid (28) Bias in AI. (z.d.). https://www.chapman.edu/ai/bias-in-ai.aspx (29) Manyika, J. (2022, 17 november). What Do We Do About the Biases in AI? Harvard Business Review. https://hbr.org/2019/10/what-do-we-do-about-the-biases-in-ai

(30) Close, K. (2022, oktober). How Tech Offers a Faster Path to Sustainability. Bcg. Geraadpleegd op 26 november 2024, van https://www.bcg.com/publications/2021/how-technology-helps-sustainability-initiatives (31) Digitalization for Sustainability. (z.d.). UNEP - UN Environment Programme. https://www.unep.org/topics/digital-transformations/digitalization-sustainability

(32) Roser, M., Ritchie, H., & Mathieu, E. (2024, 1 februari). What is Moore’s Law? Our World in Data. https://ourworldindata.org/moores-law