{"CACHEDAT":"2026-04-14 03:01:04","SLUG":"rev-01-ai-and-rising-energy-demand-QJYUpz5CB9","MARKDOWN":"# Controversy\n\n## Key ==Issue==\n\n**~~Is the ~~****~~rapid growth~~** **Are the benefits of AI worth the environmental costs tied to its energy use?**\\n→ \\n→ \n\n\n# VISUALISATION IDEAS\n\n## Commonly Misunderstood Figures (Percentages, Risks, Probabilities)\n\n| | | Evidence | Clarification or Explanation |\n|-----|-----|----------|------------------------------|\n| - | \"AI is just code, ==so== it's clean.\"
= AI is clean, ==because== it's just code.
→ | ? | Data centres and model training require physical infrastructure and large-scale power usage. |\n| ✅ | good argument | evidence supporting the good argument | |\n\n\n# Common Misrepresentations and Misperceptions\n\n## Commonly Misunderstood Figures (Percentages, Risks, Probabilities)\n\n| Misunderstood Figure | Clarification or Explanation |\n|----------------------|------------------------------|\n| \"AI is just code, so it's clean.\"
→ | Data centres and model training require physical infrastructure and large-scale power usage. |\n| \"AI uses the same electricity as regular apps.\"
→ | AI workloads, especially training, are far more energy-intensive than traditional software. |\n\n## Common Misconceptions\n\n| Misconception | Correction |\n|---------------|------------|\n| \"Using AI is always greener than manual work.\"
→ | Not necessarily — especially if the AI requires high energy or server time. |\n| \"Cloud-based AI has no environmental cost.\"
→ | Cloud servers run on electricity and generate emissions. |\n\n## Common Misinformation\n\n| Misinformation | Correction or Clarification |\n|----------------|-----------------------------|\n| AI companies are carbon-neutral.
→ | Many purchase offsets rather than reduce actual emissions. |\n| The more AI we use, the more efficient society becomes.
→ | Efficiency gains can be offset by rebound effects and rising total energy use. |\n\n\n## Key ==Issue ~~/ Question / Controversy / Debate~~==\n\n**~~Is the rapid growth~~ Are the benefits of AI worth the environmental costs tied to its energy consumption?**\n\n\n**piece of information (what you see AFTER clicking)**\n\n* 1 publication (text / video …) by… author(s) / messenger(s)\n* can include one or more YES/NO viewpoints\n\n\n**==ISSUE==**\n\n**==Viewpoint (position / stance): YES… / NO… / It's more complex…==**\\n**==→ Domain 3 (?)==**\n\nA viewpoint is the overall **answer to the issue or question**. It states where you stand. A viewpoint does not explain itself; it is a high-level position.\n\n* it normally consists of several arguments\n\n**==Argument==** \n\nAn argument is a structured justification that supports (or challenges) a viewpoint.\n\n* **Claim** – a specific, defensible statement\n* **Reasoning** – why the claim supports the viewpoint\\n==→ Domain 2==\n* **Evidence** – data, studies, observations, or well-described experience that back up the reasoning\\n==→ Domain 1==\n\n\nArgumentation → Domain 3\n\n```javascript\nIssue / Question\r\n ↓\r\nViewpoint (YES or NO)\r\n ↓\r\nArgument 1 (Claim + Reasoning + Evidence)\r\nArgument 2 (Claim + Reasoning + Evidence)\r\nArgument 3 (optional)\n```\n\n\n* innovative in SciLMi: source\n* CERS or SCERS or SCESR or SCER\n\nSteps:\n\n* > CER or only C?\n > * `D2G1O1`C: only claim → \"opinion\" unless supported by data & verified\n > * `D2G1O1H1`AWARENESS: no ER!!!\n > * `D2G1O1H2` AWARENESS: Opinions → oder checklist?\n > * CHECK → source reliable?\n > * `D2G1O1H2`CR: claim + (pseudo-)reasoning CHECK fallacies\n > * `D2G1O1H3`CE: claim + evidence\n > * + source: CHECK → source reliable\n > * + evidence: CHECK → \n >\n >\n > * – source: CHECK → subject-specific knowledge / data + FIND sources\n\n \\\n\n\n\\\n**==YES, the benefits of AI are worth the environmental costs tied to its energy consumption.==**\n\n**==Argument 1==**\n\n* **Claim:** AI can significantly reduce energy use in other sectors.\n* **Reasoning:** Efficiency gains enabled by AI can outweigh the energy it consumes.\n* **Evidence:** Peer-reviewed studies (→ D1) on AI-optimised power grids or transport systems.\n\n**==Argument 2==**\n\n* **Claim:** AI accelerates scientific research on climate mitigation.\n* **Reasoning:** Faster modelling and optimisation lead to better solutions.\n* **Evidence:** Published research outputs, climate-model improvements, etc.\n\n**==NO (Counter-Viewpoint / \"Counter-Argument\"):==** the benefits of AI are not worth the environmental costs tied to its energy consumption.\n\n**==Argument 1==**\n\n* **Claim:** Training large AI models consumes disproportionate energy.\n* **Reasoning:** Energy consumption grows faster than efficiency gains.\n* **Evidence:** Reported electricity use of large models; carbon footprint estimates.\n\n### Short rule of thumb for students\n\n* **Viewpoint** = *What do I think?*\n* **Argument** = *Why is this a good answer?*\n* **Evidence** = *How do I know?*\n\n\n***For complex issues:***\n\n* **Viewpoint = evaluated judgement**\n* **Arguments = reasons + trade-offs**\n* **Counterarguments = material to be weighed, not automatically \"the other side\"**\n\nThis mirrors **scientific argumentation**, **policy briefs**, and **ethical reasoning**.\n\n```javascript\nIssue\r\n ↓\r\nQualified Viewpoint\r\n ↓\r\nArgument 1 – Benefits (Claim + Reasoning + Evidence)\r\nArgument 2 – Costs (Claim + Reasoning + Evidence)\r\nArgument 3 – Weighing / trade-off argument\r\nArgument 4 – Conditions or safeguards\n```\n\n## Viewpoint is still singular — but *qualified*\n\nA **viewpoint does not have to be binary**.\n\nInstead of YES / NO, the viewpoint becomes a **qualified position**:\n\n* *The benefits of AI can outweigh the environmental costs **under certain conditions***\n* *The benefits currently outweigh the costs **in some domains but not others***\n* *Whether AI's benefits outweigh its costs depends on **how and where it is used***\n\nThese are **still viewpoints** because they:\n\n* Answer the question\n* Take a clear position\n* Are defensible\n\nWhat changes is that the viewpoint now includes **conditions, scope, or limits**.\n\nIn complex issues, arguments are **not all of the same type**.\n\n* a) Supporting arguments\\nThey explain **why the viewpoint holds**.\n\n\n* b) Limiting arguments\\nThey define **when / where the viewpoint does *not* hold**.\n\n\n* c) Weighing arguments (crucial for complexity)\n\n They compare competing factors and justify **why one side still prevails overall**.\n\n### Counterarguments are *internal*, not a separate viewpoint\n\nIn complex reasoning, counterarguments are often **acknowledged within the same viewpoint**, not treated as a full alternative stance.\n\nExample:\n\n> *Although AI consumes significant energy, its benefits outweigh the costs **because**…*\n\nHere:\n\n* The **counterpoint** is recognised\n* The **viewpoint remains stable**\n\nThis reflects authentic scientific and policy reasoning.\n\n\n\\\n\n\\\n**==Issue?==**\n\n* **==Are the benefits of AI worth the environmental costs tied to its energy consumption?==**\n\n**==VIEWPOINTS (= positions or stances)==**\n\n* ==YES, the benefits of AI are worth the environmental costs tied to its energy consumption.==\n* ==NO, the benefits of AI are not worth the environmental costs tied to its energy consumption.==\n\n→ When an issue is complex and cannot be reduced to a simple YES / NO, the core concepts do not change — but the structure becomes layered:\n\n* > The benefits of AI can outweigh the environmental costs under certain conditions\n >\n > The benefits currently outweigh the costs in some domains but not others\n >\n > Whether AI's benefits outweigh its costs depends on how and where it is used\n\n\nargument = claim + reasoning + evidence\n\n* **==YES: Claim==**\n * ==\"because\": **reasoning** (always supportive of viewpoint)==\n * **evidence** (statistics, personal experience…) → scientific source vs. missing scientific source / non-reliable (Check 1: scientific source?\n* **==NO: viewpoint = counterargument==**\n * ==\"because\": reasoning (always supportive of viewpoint)==\n * evidence (statistics, personal experience…) \n\n\n**Issue?**\n\n**Are the benefits of AI worth the environmental costs tied to its energy consumption?**\n\n**viewpoint** : 1 or several arguments\n\nargument = claim + reasoning + evidence\n\n* **==YES: Claim==**\n * ==\"because\": **reasoning** (always supportive of viewpoint)==\n * **evidence** (statistics, personal experience…) \n * ==\"because\": fallacy (Check 2) (here: misinformation / misconceptions)==\n * counter-evidence (?)\n* **==NO: viewpoint = counterargument==**\n * ==\"because\": reasoning (always supportive of viewpoint)==\n * evidence (statistics, personal experience…) \\n→ scientific source vs. missing scientific source / non-reliable \\n(Check 1: scientific source?\n * NO evidence → CHECK: fallacy check\n * evidence + source → CHECK reliability of the source (domain 1)\n * evidence + no source → look for sources (domain 1)\n * ==\"because\": fallacy (Check 2) (here: misinformation / misconceptions)==\n * counter-evidence (?)\n\n\n## Studies:\n\n\n\n\n\n\n\n\\+ linked studies\n\n### Main Viewpoints (+ Arguments behind these ==viewpoints==: ==fact== vs. ==fallacy /== ==opinion==)\n\n* **==YES: AI and digital systems drive innovation.==**\\n→ \n * ==BECAUSE efficiency: grid strain and rising costs (==[==iea.org==](http://iea.org)==)==\n\n ==Efficiency gains may be overstated==\n * ==\"Rebound effect\": efficiency improvements lead to increased total consumption==\n * ==Energy saved by AI optimization often enables more AI usage==\n * [https://www.cell.com/joule/fulltext/S2542-4351(23)00365-3 ](https://www.cell.com/joule/fulltext/S2542-4351(23)00365-3https://news.mit.edu/2025/how-ai-can-help-achieve-clean-energy-future-1124?utm_source=openai)\n * \"Research from MIT shows that AI is already being used to reduce energy consumption and emissions in buildings, transport, and industry, optimize the design and placement of renewable energy infrastructure, and increase the efficiency and resilience of power grids — all of which can contribute to a cleaner energy system and lower overall environmental impact\"\n * ==efficiency: AI-enabled optimization of energy systems (grids, renewables, buildings) ==\n * →\n * ==AI improves renewable energy forecasting, making solar and wind more== [==viable==](viablehttps://news.mit.edu/2023/ai-can-help-make-energy-grid-more-efficient-resilient-0327)\n * \n * AI helps economies grow (e.g. companies become more productive, etc.) \n * \\\n* **==YES: AI and digital systems drive efficiency.==**\n* ==Claim: The environmental benefits of AI can outweigh its energy consumption costs when AI is used to optimize and accelerate the transition to clean energy.==\n\n ==Reasoning: Although large AI systems consume significant energy, their ability to reduce overall emissions and improve energy efficiency in key sectors can lead to net environmental gains over time.==\n\n Evidence: \"Research from MIT shows that AI is already being used to reduce energy consumption and emissions in buildings, transport, and industry, optimize the design and placement of renewable energy infrastructure, and increase the efficiency and resilience of power grids — all of which can contribute to a cleaner energy system and lower overall environmental impact\"\n* \n* **==YES: AI and digital systems drive economic growth.==**\n* ==Claim: Artificial intelligence can drive economic growth and create significant value for companies by improving efficiency and generating large financial savings.==\n\n ==Reasoning: If AI enables companies to make operations more efficient, reduce costs, and uncover solutions that would not have been identified otherwise, then its use contributes directly to economic performance rather than being an abstract technological trend.==\n\n Evidence: \"According to Equinor, the use of artificial intelligence contributed to value creation and cost savings of 1.3 billion kroner in 2025 alone, and total value of over 3.3 billion kroner since 2020 by improving operations such as predictive maintenance, seismic data interpretation and well planning. AI is now embedded in core business processes and is seen as crucial for future profitability and energy security\"\n* **==NO: AI and digital systems does not drive economic growth.==**\n* ==Claim: Although AI is widely expected to drive economic growth and productivity, most companies have not yet seen measurable financial benefits from their AI investments.==\n\n ==Reasoning: If AI were already significantly increasing revenue or reducing costs across the economy, a large share of business leaders would report measurable results. However, many enterprises still struggle to translate AI experimentation and pilot projects into scalable value, suggesting that the link between AI and economic growth is not automatic but depends on foundational capacities such as data readiness, strategic implementation, and organizational change.==\n\n Evidence: \"According to a global PwC survey of over 4,400 CEOs, only about 29 % reported higher revenues from AI and only around 26 % reported cost reductions. Globally, only about 12 % of firms achieved both revenue growth and cost savings through AI, while a majority of CEOs (56 % worldwide and even higher in Germany) have seen no significant business results from AI initiatives so far\"\n* \n* Evidence: \"A study from the MIT Media Lab found that 95 % of AI pilot projects in U.S. companies showed no measurable positive impact on productivity or economic performance, despite significant investments; only about 5 % delivered notable revenue increases or cost savings. An additional McKinsey investigation reported that 80 % of firms using generative AI saw no significant improvements and around half abandoned their AI projects entirely\"\n* \n* **==NO: AI systems require immense energy, contributing to rising carbon emissions and resource use. → negative effects on earth outdo positive ones==**\n * * ==AI data centers consume much more power than a conventional data center.==\\n→\n * ==Data centres already account for about 1–1.5% of global electricity use and this is more than New York City…==\\n==→== [==https://www.iea.org/reports/data-centres-and-data-transmission-networks==](https://www.iea.org/reports/data-centres-and-data-transmission-networks)\n * ==AI usage is projected to triple electricity demand in data centres by 2030.== \\n==→== [==https://www.iea.org/reports/data-centres-and-data-transmission-networks==](https://www.iea.org/reports/data-centres-and-data-transmission-networks)\n * ==AI requires specialized AI chips (GPUs, TPUs), which in turn require rare earth minerals, which in turn need to be harvested, using lots of energy.==\n * ==Scientific facts: Training large AI models can emit hundreds of tonnes of CO₂.== \\n→ \n* **==YES, BECAUSE==** ==the environmental effects are unclear or not that bad because there are effective measures to reduce them.==\n * ==Mitigation through efficient siting, cooling, and operational improvements (==[==phys.org==](http://phys.org)==).==\n * supporting data / facts?\n * **==Environmental impact: of AI growth remain unclear.==**\n * ==Tech companies do not give out sufficient numbers to assess the environmental impact of AI workloads==\n\n →\n* **==NO, BECAUSE the environmental costs are even higher than most people think because the environmental costs of manufacturing the required hardware are often overlooked==**\n * ==Short hardware lifecycles create mounting e-waste==\n * ==Energy consumption: high-end hardware==\n * ==New policies mandating the disclosure of additional metrics concerning AI usage should be enforced.== →\n\n\n* **==NO, BECAUSE the future dependency on AI will create a lot of follow-up challenges that hinder innovation.==**\n * ==Excessive dependency on AI, combined with external pressure and a focus on results, creates a negative spiral that inhibits creative and deep thinking==\n * ==data? / facts?==\n * [https://www.frontiersin.org/journals/psychology/articles/10.3389/fpsyg.2025.1732837/full?utm_source=openai ](https://www.frontiersin.org/journals/psychology/articles/10.3389/fpsyg.2025.1732837/full?utm_source=openai)![](data:image/png;base64,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 \"Añadir un título al proyecto de Citavi utilizando este DOI =16x16\")\n * ==The real danger of AI in education isn't cheating — it's dependency on Big Tech algorithms==\n * \n * ==ChatGPT generally improves students' academic performance and even their motivation, while also enhancing \"higher-order thinking\" tendencies==\n * \n * ==AI-mediated dialogue fosters deeper engagement and higher-order cognitive skills.==\n * \n\n\n---\n\n# Domains of Expertise\n\n* **Environmental Science**\n * Carbon footprint calculation\n * Energy systems analysis\n * Climate change modelling\n* **Computer Science / Engineering**\n * Model architecture and efficiency\n * Cloud computing infrastructure\n * AI lifecycle energy demand\n* **Economics**\n * Cost–benefit analysis of AI adoption\n * Market incentives for green AI\n * Tech industry's energy footprint\n* **Political Science and Law**\n * Digital sustainability policy\n * AI regulation and accountability\n * Climate targets and national strategies\n* **Ethics and Philosophy**\n * Environmental justice\n * Moral responsibility of innovation\n * AI for sustainability vs. AI as a risk\n\n\n---\n\n# Main Drivers Behind the Issue\n\n* **Technological competition and innovation race**\n * AI development is fuelled by international rivalry and investor pressure.\n* **Surging demand for generative AI and automation**\n * More users rely on energy-intensive models for everyday tasks.\n* **Lack of transparency on energy usage**\n * Companies do not disclose exact emissions or energy input.\n* **Insufficient incentives for energy efficiency**\n * Green AI methods exist but are not widely adopted.\n\n→ \\n→ \n\n\n---\n\n# \n\n\n---\n\n# Parties Affected\n\n## by Impacts\n\n| Impact | Positively Affected (Individual) | Positively Affected (Organisational / Industrial) | Positively Affected (Societal) | Negatively Affected (Individual) | Negatively Affected (Organisational / Industrial) | Negatively Affected (Societal) |\n|--------|----------------------------------|---------------------------------------------------|--------------------------------|----------------------------------|---------------------------------------------------|--------------------------------|\n| Personal convenience and automation | Users of AI tools | Tech companies | Citizens with enhanced services | — | Traditional service sectors | Privacy-focused communities |\n| Data-driven decision-making | Professionals in all sectors | Enterprise and healthcare | Evidence-based policy makers | — | Small firms with low data capacity | Data justice advocates |\n| Rising energy demand and emissions | — | — | — | Citizens in climate-vulnerable areas | Renewable competitors excluded from market | Global climate resilience |\n| E-waste from hardware and upgrades | — | — | — | Low-income digital users | Repair industry | Public health infrastructure |\n| Job replacement through automation | — | Tech companies with higher margins | — | Displaced workers | Labour-intensive industries | Socio-economic equality |\n\n→ \\n→ \n\n\n---\n\n## by Potential Solutions\n\n| Potential Solution | Positively Affected (Individual) | Positively Affected (Organisational / Industrial) | Positively Affected (Societal) | Negatively Affected (Individual) | Negatively Affected (Organisational / Industrial) | Negatively Affected (Societal) |\n|--------------------|----------------------------------|---------------------------------------------------|--------------------------------|----------------------------------|---------------------------------------------------|--------------------------------|\n| Incentivise energy-efficient model design | Environmentally conscious users | AI start-ups / green tech firms | Climate and innovation policy makers | — | High-power model developers | — |\n| Mandate energy disclosure in AI development | Sustainable tech advocates | Transparency-driven platforms | Regulatory bodies | — | Secretive corporations | — |\n| Invest in renewable-powered data centres | Climate-conscious consumers | Cloud providers investing in renewables | Clean energy transition | — | Fossil-based infrastructure | — |\n| Educate users on environmental footprint | Students and professionals | Media and education sectors | Societal awareness of digital impact | Users of energy-intensive models | — | — |\n\n→ \n\n\n---\n\n# Trade-off Analysis\n\n## Individual vs. Scientific\n\n* **Convenience and productivity vs. Energy impact**\n * AI tools improve daily life but can accelerate energy consumption and emissions.\n\n## Economic vs. Environmental\n\n* **Profitability and automation vs. Ecological sustainability**\n * Firms save costs via AI while emissions and hardware demands rise.\n\n## Political vs. Scientific\n\n* **Innovation leadership vs. Carbon reduction targets**\n * Governments promote AI to stay competitive while struggling to meet climate goals.\n\n\n---\n\n# Guided Self-Reflection Prompts\n\n* **What values guide your use of digital tools and AI?**\n * Efficiency, innovation, sustainability?\n* **How do your emotions or needs shape your tech habits?**\n * Do you rely on AI for support, productivity, or connection?\n* **Have you ever questioned the environmental cost of digital convenience?**\n * What changed your thinking, if anything?\n* **What would responsible AI usage look like for you personally?**\n * Would you limit use, choose eco-friendly tools, or demand transparency?\n* **What trade-offs are you willing (or not willing) to make for sustainability?**\n * Would you accept slower tools, fewer features, or less access?\n* **==Future dependency on AI is a threat to higher education and innovation, can you find sources for this viewpoint?==**\n * ==Can you find any sources that disproves these theories?==\n\n\n---\n\n# Curricular Connections → Classroom Topics\n\n* **Computer Science (15–17)**\n * algorithm efficiency, hardware energy use, lifecycle assessments\n* **Geography / Earth Science (14–16)**\n * climate policy, digital infrastructure footprints, energy sourcing\n* **Ethics / Philosophy (15–18)**\n * justice, intergenerational responsibility, AI in the Anthropocene","HTML":"

Controversy

\n

Key Issue

\n

Is the rapid growth Are the benefits of AI worth the environmental costs tied to its energy use?\\n→ <https://www.nature.com/articles/d41586-022-01139-0>\\n→ <https://www.iea.org/reports/data-centres-and-data-transmission-networks>

\n

VISUALISATION IDEAS

\n

Commonly Misunderstood Figures (Percentages, Risks, Probabilities)

\n\n\n\n\n\n\n\n\n\n\n
EvidenceClarification or Explanation
-"AI is just code, so it's clean."
\n

= AI is clean, because it's just code. →<https://vegavid.com/blog/what-is-green-ai> | ? | Data centres and model training require physical infrastructure and large-scale power usage. | | ✅ | good argument | evidence supporting the good argument | |

\n

Common Misrepresentations and Misperceptions

\n

Commonly Misunderstood Figures (Percentages, Risks, Probabilities)

\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n
Misunderstood FigureClarification or Explanation
"AI is just code, so it's clean."
→<https://www.nature.com/articles/d41586-022-01139-0>Data centres and model training require physical infrastructure and large-scale power usage.
"AI uses the same electricity as regular apps."
→<https://www.iea.org/reports/data-centres-and-data-transmission-networks>AI workloads, especially training, are far more energy-intensive than traditional software.
\n

Common Misconceptions

\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n
MisconceptionCorrection
"Using AI is always greener than manual work."
→<https://www.iea.org/reports/data-centres-and-data-transmission-networks>Not necessarily — especially if the AI requires high energy or server time.
"Cloud-based AI has no environmental cost."
→<https://www.nature.com/articles/d41586-022-01139-0>Cloud servers run on electricity and generate emissions.
\n

Common Misinformation

\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n
MisinformationCorrection or Clarification
AI companies are carbon-neutral.
→<https://www.iea.org/reports/data-centres-and-data-transmission-networks>Many purchase offsets rather than reduce actual emissions.
The more AI we use, the more efficient society becomes.
→<https://www.nature.com/articles/d41586-022-01139-0>Efficiency gains can be offset by rebound effects and rising total energy use.
\n

Key Issue / Question / Controversy / Debate

\n

Is the rapid growth Are the benefits of AI worth the environmental costs tied to its energy consumption?

\n

piece of information (what you see AFTER clicking)

\n
    \n
  • 1 publication (text / video …) by… author(s) / messenger(s)
    • \n
  • can include one or more YES/NO viewpoints
    • \n
\n

ISSUE

\n

Viewpoint (position / stance): YES… / NO… / It's more complex…\\n→ Domain 3 (?)

\n

A viewpoint is the overall answer to the issue or question. It states where you stand. A viewpoint does not explain itself; it is a high-level position.

\n
    \n
  • it normally consists of several arguments
    • \n
\n

Argument

\n

An argument is a structured justification that supports (or challenges) a viewpoint.

\n
    \n
  • Claim – a specific, defensible statement
    • \n
  • Reasoning – why the claim supports the viewpoint\\n→ Domain 2
    • \n
  • Evidence – data, studies, observations, or well-described experience that back up the reasoning\\n→ Domain 1
    • \n
\n

Argumentation → Domain 3

\n
Issue / Question
   ↓
Viewpoint (YES or NO)
   ↓
Argument 1 (Claim + Reasoning + Evidence)
Argument 2 (Claim + Reasoning + Evidence)
Argument 3 (optional)
\n
    \n
  • innovative in SciLMi: source
    • \n
  • CERS or SCERS or SCESR or SCER
    • \n
\n

Steps:

\n
    \n
  • > CER or only C?
    • \n
\n

D2G1O1C: only claim → "opinion" unless supported by data & verified<br> D2G1O1H1AWARENESS: no ER!!!<br> D2G1O1H2 AWARENESS: Opinions → oder checklist?<br> CHECK → source reliable?<br> D2G1O1H2CR: claim + (pseudo-)reasoning CHECK fallacies<br> D2G1O1H3CE: claim + evidence<br> + source: CHECK → source reliable<br> + evidence: CHECK →<br><br><br>* – source: CHECK → subject-specific knowledge / data + FIND sources

\n

YES, the benefits of AI are worth the environmental costs tied to its energy consumption.

\n

Argument 1

\n
    \n
  • Claim: AI can significantly reduce energy use in other sectors.
    • \n
  • Reasoning: Efficiency gains enabled by AI can outweigh the energy it consumes.
    • \n
  • Evidence: Peer-reviewed studies (→ D1) on AI-optimised power grids or transport systems.
    • \n
\n

Argument 2

\n
    \n
  • Claim: AI accelerates scientific research on climate mitigation.
    • \n
  • Reasoning: Faster modelling and optimisation lead to better solutions.
    • \n
  • Evidence: Published research outputs, climate-model improvements, etc.
    • \n
\n

NO (Counter-Viewpoint / "Counter-Argument"): the benefits of AI are not worth the environmental costs tied to its energy consumption.

\n

Argument 1

\n
    \n
  • Claim: Training large AI models consumes disproportionate energy.
    • \n
  • Reasoning: Energy consumption grows faster than efficiency gains.
    • \n
  • Evidence: Reported electricity use of large models; carbon footprint estimates.
    • \n
\n

Short rule of thumb for students

\n
    \n
  • Viewpoint = What do I think?
    • \n
  • Argument = Why is this a good answer?
    • \n
  • Evidence = How do I know?
    • \n
\n

For complex issues:

\n
    \n
  • Viewpoint = evaluated judgement
    • \n
  • Arguments = reasons + trade-offs
    • \n
  • Counterarguments = material to be weighed, not automatically "the other side"
    • \n
\n

This mirrors scientific argumentation, policy briefs, and ethical reasoning.

\n
Issue
  ↓
Qualified Viewpoint
  ↓
Argument 1 – Benefits (Claim + Reasoning + Evidence)
Argument 2 – Costs (Claim + Reasoning + Evidence)
Argument 3 – Weighing / trade-off argument
Argument 4 – Conditions or safeguards
\n

Viewpoint is still singular — but qualified

\n

A viewpoint does not have to be binary.

\n

Instead of YES / NO, the viewpoint becomes a qualified position:

\n
    \n
  • The benefits of AI can outweigh the environmental costs under certain conditions
    • \n
  • The benefits currently outweigh the costs in some domains but not others
    • \n
  • Whether AI's benefits outweigh its costs depends on how and where it is used
    • \n
\n

These are still viewpoints because they:

\n
    \n
  • Answer the question
    • \n
  • Take a clear position
    • \n
  • Are defensible
    • \n
\n

What changes is that the viewpoint now includes conditions, scope, or limits.

\n

In complex issues, arguments are not all of the same type.

\n
    \n
  • a) Supporting arguments\\nThey explain why the viewpoint holds.
    • \n
\n
    \n
  • b) Limiting arguments\\nThey define when / where the viewpoint does not hold.
    • \n
\n
    \n
  • c) Weighing arguments (crucial for complexity)
    • \n
\n

They compare competing factors and justify why one side still prevails overall.

\n

Counterarguments are internal, not a separate viewpoint

\n

In complex reasoning, counterarguments are often acknowledged within the same viewpoint, not treated as a full alternative stance.

\n

Example:

\n

Although AI consumes significant energy, its benefits outweigh the costs because…

\n

Here:

\n
    \n
  • The counterpoint is recognised
    • \n
  • The viewpoint remains stable
    • \n
\n

This reflects authentic scientific and policy reasoning.

\n

Issue?

\n
    \n
  • Are the benefits of AI worth the environmental costs tied to its energy consumption?
    • \n
\n

VIEWPOINTS (= positions or stances)

\n
    \n
  • YES, the benefits of AI are worth the environmental costs tied to its energy consumption.
    • \n
  • NO, the benefits of AI are not worth the environmental costs tied to its energy consumption.
    • \n
\n

→ When an issue is complex and cannot be reduced to a simple YES / NO, the core concepts do not change — but the structure becomes layered:

\n
    \n
  • > The benefits of AI can outweigh the environmental costs under certain conditions
    • \n
\n

The benefits currently outweigh the costs in some domains but not others<br><br>Whether AI's benefits outweigh its costs depends on how and where it is used

\n

argument = claim + reasoning + evidence

\n
    \n
  • YES: Claim
    • \n
  • "because": reasoning (always supportive of viewpoint)
    • \n
  • evidence (statistics, personal experience…) → scientific source vs. missing scientific source / non-reliable (Check 1: scientific source?
    • \n
  • NO: viewpoint = counterargument
    • \n
  • "because": reasoning (always supportive of viewpoint)
    • \n
  • evidence (statistics, personal experience…)
    • \n
\n

Issue?

\n

Are the benefits of AI worth the environmental costs tied to its energy consumption?

\n

viewpoint : 1 or several arguments

\n

argument = claim + reasoning + evidence

\n
    \n
  • YES: Claim
    • \n
  • "because": reasoning (always supportive of viewpoint)
    • \n
  • evidence (statistics, personal experience…)
    • \n
  • "because": fallacy (Check 2) (here: misinformation / misconceptions)
    • \n
  • counter-evidence (?)
    • \n
  • NO: viewpoint = counterargument
    • \n
  • "because": reasoning (always supportive of viewpoint)
    • \n
  • evidence (statistics, personal experience…) \\n→ scientific source vs. missing scientific source / non-reliable \\n(Check 1: scientific source?
    • \n
  • NO evidence → CHECK: fallacy check
    • \n
  • evidence + source → CHECK reliability of the source (domain 1)
    • \n
  • evidence + no source → look for sources (domain 1)
    • \n
  • "because": fallacy (Check 2) (here: misinformation / misconceptions)
    • \n
  • counter-evidence (?)
    • \n
\n

Studies:

\n

<https://www.heise.de/news/KI-enttaeuscht-bislang-die-CEO-Hoffnungen-11147892.html>

\n

<https://www.heise.de/news/KI-steigert-Produktivitaet-aber-Unternehmen-profitieren-kaum-11067833.html>

\n

<https://taz.de/Studien-belegen-dass-der-Einsatz-von-KI-keine-Effizienzgewinne-in-Unternehmen-bringt/!6125562/>

\n

\\+ linked studies

\n

Main Viewpoints (+ Arguments behind these viewpoints: fact vs. fallacy / opinion)

\n
    \n
  • YES: AI and digital systems drive innovation.\\n→ <http://iea.org>
    • \n
  • BECAUSE efficiency: grid strain and rising costs (iea.org)
    • \n
\n

Efficiency gains may be overstated

\n
    \n
  • "Rebound effect": efficiency improvements lead to increased total consumption
    • \n
  • Energy saved by AI optimization often enables more AI usage
    • \n
  • blank\" rel=\"noopener\">https://www.cell.com/joule/fulltext/S2542-4351(23)00365-3 00365-3https://news.mit.edu/2025/how-ai-can-help-achieve-clean-energy-future-1124?utmsource=openai)<https://news.mit.edu/2025/how-ai-can-help-achieve-clean-energy-future-1124?utm_source=openai>
    • \n
  • "Research from MIT shows that AI is already being used to reduce energy consumption and emissions in buildings, transport, and industry, optimize the design and placement of renewable energy infrastructure, and increase the efficiency and resilience of power grids — all of which can contribute to a cleaner energy system and lower overall environmental impact"
    • \n
  • efficiency: AI-enabled optimization of energy systems (grids, renewables, buildings)
    • \n
  • →<http://phys.org>
    • \n
  • AI improves renewable energy forecasting, making solar and wind more viable
    • \n
  • <https://news.mit.edu/2023/ai-can-help-make-energy-grid-more-efficient-resilient-0327>
    • \n
  • AI helps economies grow (e.g. companies become more productive, etc.)
    • \n
  • \\
    • \n
  • YES: AI and digital systems drive efficiency.
    • \n
  • Claim: The environmental benefits of AI can outweigh its energy consumption costs when AI is used to optimize and accelerate the transition to clean energy.
    • \n
\n

Reasoning: Although large AI systems consume significant energy, their ability to reduce overall emissions and improve energy efficiency in key sectors can lead to net environmental gains over time.

\n

Evidence: "Research from MIT shows that AI is already being used to reduce energy consumption and emissions in buildings, transport, and industry, optimize the design and placement of renewable energy infrastructure, and increase the efficiency and resilience of power grids — all of which can contribute to a cleaner energy system and lower overall environmental impact"

\n
    \n
  • <https://news.mit.edu/2025/how-ai-can-help-achieve-clean-energy-future-1124?utm_source=openai>
    • \n
  • YES: AI and digital systems drive economic growth.
    • \n
  • Claim: Artificial intelligence can drive economic growth and create significant value for companies by improving efficiency and generating large financial savings.
    • \n
\n

Reasoning: If AI enables companies to make operations more efficient, reduce costs, and uncover solutions that would not have been identified otherwise, then its use contributes directly to economic performance rather than being an abstract technological trend.

\n

Evidence: "According to Equinor, the use of artificial intelligence contributed to value creation and cost savings of 1.3 billion kroner in 2025 alone, and total value of over 3.3 billion kroner since 2020 by improving operations such as predictive maintenance, seismic data interpretation and well planning. AI is now embedded in core business processes and is seen as crucial for future profitability and energy security"

\n
    \n
  • NO: AI and digital systems does not drive economic growth.
    • \n
  • Claim: Although AI is widely expected to drive economic growth and productivity, most companies have not yet seen measurable financial benefits from their AI investments.
    • \n
\n

Reasoning: If AI were already significantly increasing revenue or reducing costs across the economy, a large share of business leaders would report measurable results. However, many enterprises still struggle to translate AI experimentation and pilot projects into scalable value, suggesting that the link between AI and economic growth is not automatic but depends on foundational capacities such as data readiness, strategic implementation, and organizational change.

\n

Evidence: "According to a global PwC survey of over 4,400 CEOs, only about 29 % reported higher revenues from AI and only around 26 % reported cost reductions. Globally, only about 12 % of firms achieved both revenue growth and cost savings through AI, while a majority of CEOs (56 % worldwide and even higher in Germany) have seen no significant business results from AI initiatives so far"

\n
    \n
  • <https://www.heise.de/news/KI-enttaeuscht-bislang-die-CEO-Hoffnungen-11147892.html>
    • \n
  • Evidence: "A study from the MIT Media Lab found that 95 % of AI pilot projects in U.S. companies showed no measurable positive impact on productivity or economic performance, despite significant investments; only about 5 % delivered notable revenue increases or cost savings. An additional McKinsey investigation reported that 80 % of firms using generative AI saw no significant improvements and around half abandoned their AI projects entirely"
    • \n
  • <https://taz.de/Studien-belegen-dass-der-Einsatz-von-KI-keine-Effizienzgewinne-in-Unternehmen-bringt/!6125562/>
    • \n
  • NO: AI systems require immense energy, contributing to rising carbon emissions and resource use. → negative effects on earth outdo positive ones
    • \n
  • * AI data centers consume much more power than a conventional data center.\\n→<https://www.sciencedirect.com/science/article/pii/S2542435124003477>
    • \n
  • Data centres already account for about 1–1.5% of global electricity use and this is more than New York City…\\n→ https://www.iea.org/reports/data-centres-and-data-transmission-networks
    • \n
  • AI usage is projected to triple electricity demand in data centres by 2030. \\n→ https://www.iea.org/reports/data-centres-and-data-transmission-networks
    • \n
  • AI requires specialized AI chips (GPUs, TPUs), which in turn require rare earth minerals, which in turn need to be harvested, using lots of energy.
    • \n
  • Scientific facts: Training large AI models can emit hundreds of tonnes of CO₂. \\n→ <https://www.nature.com/articles/d41586-022-01139-0>
    • \n
  • YES, BECAUSE the environmental effects are unclear or not that bad because there are effective measures to reduce them.
    • \n
  • Mitigation through efficient siting, cooling, and operational improvements (phys.org).
    • \n
  • supporting data / facts?
    • \n
  • Environmental impact: of AI growth remain unclear.
    • \n
  • Tech companies do not give out sufficient numbers to assess the environmental impact of AI workloads
    • \n
\n

→<https://www.sciencedirect.com/science/article/pii/S2666389925002788>

\n
    \n
  • NO, BECAUSE the environmental costs are even higher than most people think because the environmental costs of manufacturing the required hardware are often overlooked
    • \n
  • Short hardware lifecycles create mounting e-waste
    • \n
  • Energy consumption: high-end hardware
    • \n
  • New policies mandating the disclosure of additional metrics concerning AI usage should be enforced. →<https://www.sciencedirect.com/science/article/pii/S2666389925002788 ](https://www.sciencedirect.com/science/article/pii/S2666389925002788)<https://www.cell.com/joule/fulltext/S2542-4351(23)00365-3>
    • \n
\n
    \n
  • NO, BECAUSE the future dependency on AI will create a lot of follow-up challenges that hinder innovation.
    • \n
  • Excessive dependency on AI, combined with external pressure and a focus on results, creates a negative spiral that inhibits creative and deep thinking
    • \n
  • data? / facts?
    • \n
  • source=openai\" target=\"blank\" rel=\"noopener\">https://www.frontiersin.org/journals/psychology/articles/10.3389/fpsyg.2025.1732837/full?utm_source=openai \"\"
    • \n
  • The real danger of AI in education isn't cheating — it's dependency on Big Tech algorithms
    • \n
  • <https://www.businessinsider.com/ai-is-handing-control-of-knowledge-to-big-tech-professor-2025-10?utm_source=openai>
    • \n
  • ChatGPT generally improves students' academic performance and even their motivation, while also enhancing "higher-order thinking" tendencies
    • \n
  • <https://www.sciencedirect.com/science/article/pii/S0360131524002380>
    • \n
  • AI-mediated dialogue fosters deeper engagement and higher-order cognitive skills.
    • \n
  • <https://arxiv.org/abs/2509.16262?utm_source=openai>
    • \n
\n
\n

Domains of Expertise

\n
    \n
  • Environmental Science
    • \n
  • Carbon footprint calculation
    • \n
  • Energy systems analysis
    • \n
  • Climate change modelling
    • \n
  • Computer Science / Engineering
    • \n
  • Model architecture and efficiency
    • \n
  • Cloud computing infrastructure
    • \n
  • AI lifecycle energy demand
    • \n
  • Economics
    • \n
  • Cost–benefit analysis of AI adoption
    • \n
  • Market incentives for green AI
    • \n
  • Tech industry's energy footprint
    • \n
  • Political Science and Law
    • \n
  • Digital sustainability policy
    • \n
  • AI regulation and accountability
    • \n
  • Climate targets and national strategies
    • \n
  • Ethics and Philosophy
    • \n
  • Environmental justice
    • \n
  • Moral responsibility of innovation
    • \n
  • AI for sustainability vs. AI as a risk
    • \n
\n
\n

Main Drivers Behind the Issue

\n
    \n
  • Technological competition and innovation race
    • \n
  • AI development is fuelled by international rivalry and investor pressure.
    • \n
  • Surging demand for generative AI and automation
    • \n
  • More users rely on energy-intensive models for everyday tasks.
    • \n
  • Lack of transparency on energy usage
    • \n
  • Companies do not disclose exact emissions or energy input.
    • \n
  • Insufficient incentives for energy efficiency
    • \n
  • Green AI methods exist but are not widely adopted.
    • \n
\n

→ <https://www.nature.com/articles/d41586-022-01139-0>\\n→ <https://www.iea.org/reports/data-centres-and-data-transmission-networks>

\n
\n
\n

Parties Affected

\n

by Impacts

\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n
ImpactPositively Affected (Individual)Positively Affected (Organisational / Industrial)Positively Affected (Societal)Negatively Affected (Individual)Negatively Affected (Organisational / Industrial)Negatively Affected (Societal)
Personal convenience and automationUsers of AI toolsTech companiesCitizens with enhanced services—Traditional service sectorsPrivacy-focused communities
Data-driven decision-makingProfessionals in all sectorsEnterprise and healthcareEvidence-based policy makers—Small firms with low data capacityData justice advocates
Rising energy demand and emissions———Citizens in climate-vulnerable areasRenewable competitors excluded from marketGlobal climate resilience
E-waste from hardware and upgrades———Low-income digital usersRepair industryPublic health infrastructure
Job replacement through automation—Tech companies with higher margins—Displaced workersLabour-intensive industriesSocio-economic equality
\n

→ <https://www.nature.com/articles/d41586-022-01139-0>\\n→ <https://www.iea.org/reports/data-centres-and-data-transmission-networks>

\n
\n

by Potential Solutions

\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n
Potential SolutionPositively Affected (Individual)Positively Affected (Organisational / Industrial)Positively Affected (Societal)Negatively Affected (Individual)Negatively Affected (Organisational / Industrial)Negatively Affected (Societal)
Incentivise energy-efficient model designEnvironmentally conscious usersAI start-ups / green tech firmsClimate and innovation policy makers—High-power model developers—
Mandate energy disclosure in AI developmentSustainable tech advocatesTransparency-driven platformsRegulatory bodies—Secretive corporations—
Invest in renewable-powered data centresClimate-conscious consumersCloud providers investing in renewablesClean energy transition—Fossil-based infrastructure—
Educate users on environmental footprintStudents and professionalsMedia and education sectorsSocietal awareness of digital impactUsers of energy-intensive models——
\n

→ <https://www.iea.org/reports/data-centres-and-data-transmission-networks>

\n
\n

Trade-off Analysis

\n

Individual vs. Scientific

\n
    \n
  • Convenience and productivity vs. Energy impact
    • \n
  • AI tools improve daily life but can accelerate energy consumption and emissions.
    • \n
\n

Economic vs. Environmental

\n
    \n
  • Profitability and automation vs. Ecological sustainability
    • \n
  • Firms save costs via AI while emissions and hardware demands rise.
    • \n
\n

Political vs. Scientific

\n
    \n
  • Innovation leadership vs. Carbon reduction targets
    • \n
  • Governments promote AI to stay competitive while struggling to meet climate goals.
    • \n
\n
\n

Guided Self-Reflection Prompts

\n
    \n
  • What values guide your use of digital tools and AI?
    • \n
  • Efficiency, innovation, sustainability?
    • \n
  • How do your emotions or needs shape your tech habits?
    • \n
  • Do you rely on AI for support, productivity, or connection?
    • \n
  • Have you ever questioned the environmental cost of digital convenience?
    • \n
  • What changed your thinking, if anything?
    • \n
  • What would responsible AI usage look like for you personally?
    • \n
  • Would you limit use, choose eco-friendly tools, or demand transparency?
    • \n
  • What trade-offs are you willing (or not willing) to make for sustainability?
    • \n
  • Would you accept slower tools, fewer features, or less access?
    • \n
  • Future dependency on AI is a threat to higher education and innovation, can you find sources for this viewpoint?
    • \n
  • Can you find any sources that disproves these theories?
    • \n
\n
\n

Curricular Connections → Classroom Topics

\n
    \n
  • Computer Science (15–17)
    • \n
  • algorithm efficiency, hardware energy use, lifecycle assessments
    • \n
  • Geography / Earth Science (14–16)
    • \n
  • climate policy, digital infrastructure footprints, energy sourcing
    • \n
  • Ethics / Philosophy (15–18)
    • \n
  • justice, intergenerational responsibility, AI in the Anthropocene
    • \n
","UPDATEDAT":"2026-02-26T13:29:54.938Z","ID":"cb78351c-eef6-4d8b-8d40-d8edb51c3638","TITLE":"REV 01 - AI and rising energy demand"}