I’m still in two minds about AI. Sure, it has revolutionised system automation, data analysis and data processing, and it’s at the forefront of customer service and user experience. But when it comes to the creative arts, AI is several shades of grey. While it’s democratised the creative process (or, more accurately, the ability to produce something from a set of instructions), its ability to mimic other creators’ styles – and artistic creations is problematic.<\/p>\n\n\n\n
I am one of the many authors whose work was used – without authority, attribution or payment – to train an AI Large Language Model. It was a two-stage crime. Firstly, pirated (i.e. stolen) books were made available on the LibGen website for free. Secondly, Meta scraped that ‘library’ to train its AI, arguing that said information was already freely available.<\/p>\n\n\n\n
You can search for the authors and works affected by Meta’s AI LLM here<\/a>.<\/strong><\/p>\n\n\n\n There’s another aspect to AI that doesn’t get much airplay. AI isn’t free. Not really. There is an environmental impact. Cooling methods:<\/strong> If you’re going to use AI may I suggest you also donate to WaterAid<\/strong><\/a>.<\/p>\n\n\n\n <\/p>\n","protected":false},"excerpt":{"rendered":" I’m still in two minds about AI. Sure, it has revolutionised system automation, data analysis and data processing, and it’s at the forefront of customer service and user experience. But when it comes to the creative arts, AI is several … Continued<\/a><\/p>\n","protected":false},"author":1,"featured_media":190,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"kt_blocks_editor_width":"","footnotes":""},"categories":[1,10],"tags":[27,19,31,30,29,34,28,33,32],"class_list":["post-171","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-derek_thompson","category-writing-and-publishing","tag-ai","tag-author","tag-chatgpt","tag-creative-arts","tag-large-language-model","tag-libgen","tag-llm","tag-meta-ai","tag-open-ai"],"_links":{"self":[{"href":"https:\/\/dpnwordpress.org\/derekthompson\/wp-json\/wp\/v2\/posts\/171","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/dpnwordpress.org\/derekthompson\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/dpnwordpress.org\/derekthompson\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/dpnwordpress.org\/derekthompson\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/dpnwordpress.org\/derekthompson\/wp-json\/wp\/v2\/comments?post=171"}],"version-history":[{"count":8,"href":"https:\/\/dpnwordpress.org\/derekthompson\/wp-json\/wp\/v2\/posts\/171\/revisions"}],"predecessor-version":[{"id":189,"href":"https:\/\/dpnwordpress.org\/derekthompson\/wp-json\/wp\/v2\/posts\/171\/revisions\/189"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/dpnwordpress.org\/derekthompson\/wp-json\/wp\/v2\/media\/190"}],"wp:attachment":[{"href":"https:\/\/dpnwordpress.org\/derekthompson\/wp-json\/wp\/v2\/media?parent=171"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/dpnwordpress.org\/derekthompson\/wp-json\/wp\/v2\/categories?post=171"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/dpnwordpress.org\/derekthompson\/wp-json\/wp\/v2\/tags?post=171"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}
The amount of water used to cool data centres per AI question varies depending on several factors, including the type of AI model, data centres location and cooling systems.
Water usage estimates:<\/strong>
– ChatGPT query: approximately 0.5 litres (17 oz) of water for 100-word email generation or 2 litres (\u00bd gallon) for 10 to 50 queries.
– Teaspoon of water:* about 1 teaspoon of water per query, according to OpenAI CEO Sam Altman.
<\/p>\n\n\n\n
– Evaporative cooling is the most common method, where hot air passes over water, causing evaporation that removes heat, but results in significant water loss.
– Air cooling uses fans and air circulation to dissipate heat, requiring less water but less efficient for high-density workloads.
– Immersion cooling is a newer approach where servers are submerged in a special liquid that absorbs heat, drastically reducing water use and improving efficiency.
Data centre water usage:<\/strong>
– Microsoft’s Iowa data centres used over 11 million gallons of water on peak summer days in 2022, accounting for about 6% of the district’s total water consumption.
– Google’s water usage was over 5.6 billion gallons of water in one year, with plans to form a data centre in Arizona that will work on “air-cooled technology” to mitigate water shortage.
SOURCE: Meta AI. (I kid you not.)<\/p>\n\n\n\n