> ## Content Index > Fetch the complete content index at: https://www.moderndruidmagazine.com/llms.txt > Use this file to discover other available public pages before exploring further. # What To Do About AI? - URL: https://www.moderndruidmagazine.com/whats-a-druid-to-do-about-ai/ - Published: 2026-09-22T23:49:00.000Z - Updated: 2026-09-23T00:47:12.000Z - Description: How to navigate the increasingly complex threat of AI - Author: Morgana Brennan - Tags: SCIENTIA A forge requires fire. A mill requires water. Words sent instantaneously across the ocean via a cable still require a wire, a satellite, a roomful of humming machines. So does artificial intelligence, because it is fire and water made electric, and as such, requires tending rather than being put out of concern or without thought thrown onto the fire. How a Druid's approach to power is always through discernment: knowing what something costs; who has it in good care; and how one's own hands might keep it in proper alignment with the earth. AI belongs to that realm of inquiry today. Four suggestions follow. ### **Backing the Water-Keepers** The heat generated by AI is tangible and present. The sheer number of processors at work in every big data center means that each such place, effectively, has a whole small city's worth of computer power, all of which must be cooled or burned up. Global data-center energy usage was at 415 terawatt-hours in 2024 and on track to hit around 945 terawatt-hours by 2030, while AI-specific demands jumped about 50% in 2025\. Cooling accounts for much of that, and water for much of the cooling, making the companies that do it well valuable to know. A reader whose capital she would like to see moving towards solutions has some real, traded options to investigate, listed below. This is not financial advice. All of these companies deserve some research of her own – reading the filings, investigating how they define their efficiency, and possibly a consultation with a financial advisor. ![](https://storage.ghost.io/c/e1/2b/e12b16e4-d1c4-4075-97bb-b7641c509179/content/images/2026/09/Screenshot-2026-09-22-at-12.22.34---AM.png) Microsoft's own water usage efficiency increased by 39% between 2021 and 2025, proving that this is a solvable problem and that the market rewards it. ### **Reaching the Web Without Waking the Machine** Nowadays, AI is omnipresent in the internet space. Even the most trivial of queries, such as searching for a phone number, an oven temperature conversion for the recipe or the capital of a country – a trivial task requiring only a simple fact to be retrieved from the web – now takes an unnecessarily large amount of computing resources as it passes through the large language model. Those search engines, who still make the AI-generated summaries optional or unavailable, allow a reader to save on her electricity expenses as well as the carbon emissions, gathered here for easy reference: ![](https://storage.ghost.io/c/e1/2b/e12b16e4-d1c4-4075-97bb-b7641c509179/content/images/2026/09/Screenshot-2026-09-22-at-12.34.56---AM.png) Set one of these as a default browser, and use a search bar or a direct URL if the destination is known – this will keep the fires burning only where a question requires reasoning rather than retrieval. ### **Tending AI as a Careful Flame** When a task truly requires AI input, there are real ways to do it in a light fashion. Text generation is on the more frugal side of the spectrum. A modest request consumes roughly the energy of a few seconds spent on an electric bike. The image generation costs significantly more and the video generation takes exponentially more energy. The researchers from MIT Technology Review discovered that the short, high-quality AI video clip could consume more than 700 times the energy compared to one high-quality image! Choosing text instead of an image, and image instead of a video, these sorts of choices are the largest lever we hold. And making one complete request rather than several exchanges cuts on the energy use since each subsequent exchange requires recontextualization of the information again from the very beginning. Using a 'standard' or lighter model version for routine tasks and reserving the large, deep models for solving difficult problems also makes a significant difference, because the largest models consume many times the energy compared to the properly sized model for the simple task. For those readers who seek an off-grid solution, small open-weight models can be run locally on a laptop through tools like **Ollama or LM Studio**, generating text and images without engaging the data center at all. This is the only way to localize the AI fully. But note, there is a misperception that a desktop application of a cloud-based assistant reduces the drag, it does not, it only serves as a convenient window to it, but the computation happens in a data center somewhere else, regardless of the access point. Being aware of this is part of discernment. ### **Backing the Keepers of the Threshold** The question the most important to ask of any emerging power is always "Who stands at the threshold of it, and what do they allow to pass through?" Anthropic offered an explicit answer to this question earlier this year. When the US government asked to use its AI technology for domestic surveillance of its citizens, the company refused. In response, the Department of Defense labeled Anthropic a “supply chain risk.” The company fought back, and on March 24, 2026, the court issued a preliminary injunction against the government, ruling that this decision was punitive. The Electronic Frontier Foundation and the Foundation for Individual Rights and Expression submitted a brief in support. Containment of any emerging technology means a company prepared to suffer real losses for the sake of closing a threshold. However, there is another side of the story. On September 9, 2026, researcher Jacob Coxon left Anthropic and posted his resignation publicly, saying that “They are racing straight to self-improving superintelligence and gambling with our lives”, referring both to Anthropic and to his former employer, OpenAI. His fellow researchers from Anthropic, Evan Hubinger and Anna Wang, supported his viewpoint. Dario Amodei responded quickly with an essay called “We Must Pace the Frontier”, calling for independent evaluations of the frontier projects, agreements on the development pace between the industries and international cooperation. Anthropic's record contains contradictions as well – back in February 2026, Anthropic changed its founding Responsible Scaling Policy to give itself an exception to develop even if the lead over its competitors was small enough, which does not really fit into its previous positioning as the safe alternative. > Thus, we find ourselves asking the question: who to trust? No-one. Both of these sides are true at the same time, and this is precisely the discernment that the practice requires. A company can defend the threshold from one attack but hesitate before another under competitive pressure – which is exactly why the vigilance must be a continuous practice rather than an isolated verdict. [**The Future of Life Institute's AI Safety Index** ](https://futureoflife.org/ai-safety-index-summer-2026/?ref=moderndruidmagazine.com)offers precisely such a tool to a reader. Its Summer 2026 assessment evaluated nine major AI companies on thirty-seven metrics of risk assessment, current harms, safety framework, existential safety, governance and transparency. Anthropic scored the best mark in this field (C+), followed by OpenAI and Google DeepMind, while xAI, DeepSeek and Mistral received an F for minimal or even totally absent safety infrastructure. Investigating this index before using the particular AI services and before choosing the companies to buy from as customers and as stockholders turns an abstraction into a concrete practice, which should be repeated every time the index is updated, not memorized. ![](https://storage.ghost.io/c/e1/2b/e12b16e4-d1c4-4075-97bb-b7641c509179/content/images/2026/09/Screenshot-2026-09-22-at-2.58.17---PM.png) AI Safety Index, Summer 2026 | futureoflife.org ### **Closing the Circle** None of these asks anything from a reader, neither acceptance nor rejection of AI. It asks for the same stewardship towards any power invited to the grove: to know its cost, to support those tending it responsibly, to use it intentionally, and to guard continuously those who stand at its gates. Choosing even one of these four practices this season: researching a stock, switching the search engine, avoiding the video prompts in favor of texts, keeping the safety index bookmarked, that is enough to begin to transform the relationship from consumption to right relationship. --- ### **Sources** - [Powering AI: Barclays Names Top Data Center Cooling Stocks to Own — Investing.com](https://www.investing.com/news/stock-market-news/powering-ai-barclays-names-top-data-center-cooling-stocks-to-own-93CH-4603432?ref=moderndruidmagazine.com) - [11 AI Data Center Cooling Stocks to Own After the Ecolab-CoolIT Deal — MarketWise](https://marketwise.com/investing/ai-data-center-cooling-stocks-ecolab-coolit/?ref=moderndruidmagazine.com) - [AI Energy Statistics (2026) — VoxBooster](https://voxbooster.com/blog/ai-energy-statistics-2026/?ref=moderndruidmagazine.com) - [We did the math on AI's energy footprint — MIT Technology Review](https://www.technologyreview.com/2025/05/20/1116327/ai-energy-usage-climate-footprint-big-tech/?ref=moderndruidmagazine.com) - [Which Search Engine Doesn't Use AI? 8 Picks for 2026 — Ban the Bots](https://www.banthebots.org/explainers/search-engines-without-ai?ref=moderndruidmagazine.com) - [AI Safety Index — Summer 2026 — Future of Life Institute](https://futureoflife.org/ai-safety-index-summer-2026/?ref=moderndruidmagazine.com) - [The Government Must Not Force Companies to Participate in AI-powered Surveillance — EFF](https://www.eff.org/deeplinks/2026/03/government-must-not-force-companies-participate-ai-powered-surveillance?ref=moderndruidmagazine.com) - [“I resigned from Anthropic” — AI whistleblower — TheGrio](https://thegrio.com/2026/09/13/ai-whistleblower-tech-leaders-slowdown/?ref=moderndruidmagazine.com) - [“Hit the brakes”: Anthropic warns the race to superintelligence — Genetic Literacy Project](https://geneticliteracyproject.org/2026/09/14/hit-the-brakes-anthropic-warns-the-race-to-superintelligence-could-spawn-internet-seizing-ai-swarms-and-empower-authoritarians-heres-how-to-stop-it/?ref=moderndruidmagazine.com) - [Anthropic Drops Hallmark Safety Pledge in Race With AI Peers — Bloomberg](https://www.bloomberg.com/news/articles/2026-02-25/anthropic-adds-caveat-to-ai-safety-policy-in-race-against-rivals?ref=moderndruidmagazine.com) --- ***Morgana Brennan, MBA*** *Morgana Brennan is a practicing Druid, Ogham Practitioner and Chief Magic Maker at Modern Druid, the gallery, apothecary, healing space, and event studio she founded at 60 South Broadway in Nyack, New York in 2021\.* *She holds a BSc in Economics and Business, a dual MBA in Management Systems and Information & Communication Systems, and an MIT certificate in Dynamic Work Design, and she brings her professional wizardry to other of her businesses consulting with and supporting independent healthcare practices.* *Raised in Ireland and rooted in the Druid traditions of her homeland, she trained through OBOD and the Irish Pagan School. From that work she created her own Ogham intention oils and Morgana's Window, a divination practice read across a custom Fionn's Window table.* *She gathers community the way an oak gathers birds. She founded DruidFest, hosts a monthly Druid Roundtable, opens Threshold Gatherings alongside fellow practitioners, and co-hosts the Modern Druid Podcast with her husband, Eric. The Modern Druid Promptuarium is her latest act of conjuring: a living periodical of doorways, each opening onto a different corner of the sacred.* *Both lives run on the same magic. She takes what is tangled and complex, whether in the business of healthcare or an ancient tree alphabet, and turns it into something people can use. Every word, in her view, is abracadabra: as it is spoken, so it becomes.* [Modern DruidA purveyor beautiful, mindful, spiritual products and services for the intentional, curious soul. Modern Druid offers spiritual tools to those seeking tangible ways to engage with Spirit, through divination, intention and art. Anchoring our curation is a unique line of Ogham inspired essential oil blends divined to allow for a more tangible way to work with these ancient Irish spiritual symbols.![](https://storage.ghost.io/c/e1/2b/e12b16e4-d1c4-4075-97bb-b7641c509179/content/images/icon/rs-w-64-h-64-m-b3c8c8d2-d0c3-4ea1-999a-8d980b247476.png)Modern DruidModern Druid![](https://storage.ghost.io/c/e1/2b/e12b16e4-d1c4-4075-97bb-b7641c509179/content/images/thumbnail/ModernDruid_Logo_Gold_Rev-d98efe76-c078-42af-a64f-9b4088fda605.png)](https://modern-druid.com/?ref=moderndruidmagazine.com)