# Part 6: How do LLMs impact society? Source: https://www.youtube.com/watch?v=5MQIJTDOBrM Recap page: https://rapidrecap.app/video/5MQIJTDOBrM Generated: 2026-01-23T05:25:00.052+00:00 --- ## Quick Overview The overall environmental impact of LLMs is projected to be much worse across energy consumption, water footprint, and carbon emissions based on current growth scenarios, despite the hype suggesting otherwise, while experts and developers consistently overestimate the productivity gains from AI tools in real-world tasks. **Key Points:** - The environmental impact of LLMs is projected to be much worse by 2030 across energy consumption (up to 444.9 TWh in the highest case), water footprint (up to 1124.5 million m³), and carbon emissions (up to 80 Mt), according to projections from a recent study. - Data centers are increasingly located in water-stressed areas, with large data centers consuming up to 5 million gallons of water per day, equivalent to the water use of a town of 10,000 to 50,000 people. - A survey of 48,000 people across 47 countries shows that while 73% use general-purpose generative AI tools, 48% of employees report uploading company information into public AI tools, raising liability concerns. - Actual measurements from a randomized controlled trial with developers show that AI usage slows down task completion time by about 20% (contrary to the 30-40% speedup often estimated by experts and developers). - The work impact is characterized by two claim types: 'It helps me do something I couldn't otherwise do' (more beneficial for novices) versus 'It has the potential to revolutionize something' (often marketing hype). - Learning to use LLMs ('prompt engineering') is insufficient for experts, who show no strong evidence of improvement over doing the work themselves, and may hinder skill development for novices by creating dependency on subpar AI output. - The current AI business model risks 'enshittification' and is propelled by circular investment, where companies buy from each other hoping not to be the last one holding the bag when the bubble bursts or prices rise significantly. ![Screenshot at 40:24: The speaker transitions to the environmental impact, stating that the environmental impact of LLMs is 'terrible,' using an image of a crystal ball reflecting a dry, cracked landscape with a waterfall, symbolizing a distorted or worrying environmental future.](https://ss.rapidrecap.app/screens/5MQIJTDOBrM/00-40-24.jpg) **Context:** This presentation, identified as Part 6 of an 'All about AI' series focusing on societal impacts, transitions into an analysis of the economic and environmental consequences of Large Language Models (LLMs) and AI. The speaker moves from discussing job impacts (where LLMs might eliminate jobs or hinder skill development) to economic fears of an AI bubble, evidenced by massive capital expenditures not matched by current revenues, and finally addresses the significant environmental footprint of data centers, particularly concerning water consumption and carbon emissions. ## Detailed Analysis The presentation concludes the societal impact series by focusing on the economic and environmental consequences of LLMs, arguing that the benefits are overstated and the risks are significant. Economically, the presentation highlights fears of an AI bubble, supported by data showing massive capital expenditures by tech giants (Meta, Google, Microsoft, Amazon) that far outpace current revenues, suggesting unsustainable growth reliant on investor equity. The speaker notes that the supposed productivity gains from AI are often exaggerated; a study on software developers found that AI usage actually slowed down task completion time by about 20% for those using AI, compared to the control group, suggesting real-world utility lags behind hype. Furthermore, the presentation addresses the danger of 'enshittification'—where products degrade over time to maximize short-term profits—and the risk of dependency on proprietary tools that lack transparency, leading to potential liability issues. Environmentally, the impact is severe: AI data centers are consuming enormous amounts of electricity (4.4% of US total in 2023, projected to triple by 2028) and water (up to 5 million gallons per day for large centers, often in water-stressed regions), leading to increased carbon emissions. The speaker concludes by suggesting that simply learning 'prompt engineering' is not a universal solution, as it may only help novices while experts see little benefit, and that the entire ecosystem might be driven by unsustainable economic incentives. ### LLM Impacts on Work - Two claim types exist: 'It helps me do something I couldn't otherwise do' (often for novices) and 'It has the potential to revolutionize something' (often hype); experts show no strong evidence of benefit over doing work themselves, and reliance on AI may hinder skill development. ### LLM Impacts on Work - If LLMs improve productivity, consequences include job elimination and destruction of the pathway for higher expertise development, as juniors rely on AI instead of learning fundamentals. ### Economic Impacts of LLMs - Fears of an AI bubble are grounded because massive capital expenditures by tech companies ($30B+ in generative AI investment alone) are not matched by current revenues, indicating unsustainable growth propped up by investor equity. ### Economic Impacts of LLMs - A study showed that developers using AI were actually slowed down (observed result around -20% time change) compared to expert forecasts (around -40%), indicating overestimation of productivity gains. ### Business model - ensittification: This process, also called crapification or platform decay, involves vendors creating high-quality offerings to attract users, then degrading services to maximize short-term shareholder profits, locking users into dependency. ### Environmental Context - AI data centers consume massive energy (4.4% of US electricity in 2023, projected 3x by 2028) and water (up to 5 million gallons/day per large center) often in water-stressed areas, leading to escalating carbon emissions. ### Next video - Part 7: My recommendations: The speaker concludes this section by pointing towards the final video, which will offer recommendations. ![Screenshot at 00:23: The presentation begins with the title slide for Part 6: 'How do LLMs impact society?', showing a digital globe with radiating lines, setting the stage for a broad discussion on societal effects.](https://ss.rapidrecap.app/screens/5MQIJTDOBrM/00-00-23.jpg) ![Screenshot at 04:38: A slide categorizing claims about LLM impacts on work into two types: 'It helps me do something I couldn't otherwise do' and 'It has the potential to revolutionize something', with sub-questions for each category.](https://ss.rapidrecap.app/screens/5MQIJTDOBrM/00-04-38.jpg) ![Screenshot at 07:37: A flowchart illustrating the analysis of LLM productivity, leading to the conclusion that if productivity gains are real, the consequences are job elimination and destruction of the pathway for higher expertise development.](https://ss.rapidrecap.app/screens/5MQIJTDOBrM/00-07-37.jpg) ![Screenshot at 11:18: A bar chart showing the frequency of intentional AI use at work, broken down by Global, Advanced Economy, and Emerging Economy respondents, indicating variation in adoption rates.](https://ss.rapidrecap.app/screens/5MQIJTDOBrM/00-11-18.jpg) ![Screenshot at 33:20: A slide titled 'LLM impacts on work' featuring a dark, evocative image of a person leaning against a wall in distress, overlaid with the warning: 'We run the danger of embedding tools we don't control into our processes, and then becoming dependent on them.'](https://ss.rapidrecap.app/screens/5MQIJTDOBrM/00-33-20.jpg)