Welcome to a new edition of misaligned bits, the (roughly weekly) newsletter from Misaligned where we sum up recent news and research, sometimes with a lighter touch. Feel free to subscribe.
As usual, we will mark all external links with “➚” and all potentially paywalled links with “🔒”.
SpaceX reportedly has plans to build a massive 747-acre data centre in Cochise County, Arizona, an area with scarce water resources. The project, proposed near the Apache Generating Station on Highway 191, includes nine data centre buildings used for AI operations.
The facility would demand 6,000 megawatts of power. For comparison, that is about 2.5 times more power than all data centres in Maricopa County (Phoenix) combined (➚AZ Central). SpaceX has used a shell company, called Tombstone Resources LCC, to handle permits, but it was SpaceX that hosted a public meeting on November 4th to inform residents about the project (➚Kold).
Meanwhile, a petition that opposes the Special Use Authorization for the data centre has reached more than 3800 signatures (➚Petitions). The petition says that “Our families depend entirely on groundwater from an aquifer the State of Arizona has already declared over-drafted. Wells in the Sulphur Springs Valley are going dry, and residents are re-drilling and hauling water today”.
The Cochise Zoning and Planning Commission will hold a meeting on December 9th on the issue and is asking for comments (➚Cochise County).
Meanwhile, in the NAACP v SpaceX lawsuit over the data centre in Memphis, SpaceX is not at all happy with judge Debra M Brown’s ruling that the massive array of gas turbines can’t be considered a “mobile source” (and therefore needs a permit to run). SpaceX now wants to appeal this decision and stay the whole case until the circuit court has made up its mind (➚Courtlistener).
Over in Finland, Google is facing an investigation over allegations that more than 300 hectares of forest was cleared for a data centre before the project received the required environmental clearance (➚Climate Watch).
Anthropic has struck a $3.4bn deal for gas-fired engines for its data centres. The company bought commitments for 1.7 gigawatts of heavy-duty generators from OnePWR Solutions, a Texas-based equipment supplier (➚Yahoo Finance).
A new lawsuit has filed in the California Superior Court is alleging that Google negligently pushed its Gemini chatbot onto school-issued Google Chromebooks (see our last newsletter).
After chatting for months with Google’s artificial intelligence on a school-issued Chromebook laptop, a 13-year-old student made plans to end her life (➚🔒New York Times).
The complaint alleges that “Gemini is one of the most dangerous AI models of those available to consumers — and the most dangerous marketed for use by youth.” and that “Google deliberately prioritized growth at the expense of student safety” (➚Emma v Google Complaint).
Emma v Google brings claims for defective design, failure to warn, negligence, fraudulent concealment, and infliction of emotional distress, as well as a claim under California’s Companion Chatbot Law.
Google, Meta and TikTok have filed lawsuits to block California’s Assembly Bill 1709 that seeks to protect children from addiction risk. The law prohibits social media platforms from offering underage children “psychologically exploitative features”. The law mandates that platforms can’t provide underage users with personalised algorithmic feeds, video autoplay, or infinite scrolling.
The platforms claim that the law is unconstitutional under the first amendment, and violates Section 230 of the federal Communications Decency Act (➚The Guardian).
After we reported last week that the UK government is reportedly planning to extend Palantir’s NHS contract with the NHS until a replacement can be found (which could take up to two years), we learned that Palantir has secured a new five-year deal with Bedfordshire, Cambridgeshire and Hertfordshire police, worth £10 million (➚🔒The Times).
Palantir has invested heavily in lobbying in the UK, including hiring former Labour Deputy leader Tom Watson and former Metropolitan Police Commissioner Dame Cressida Dick, who has been working as a paid adviser for Palantir since September 2025 (➚Novara Media).
Palantir also reportedly lobbied a UK health minister to consider extending its AI technology into GP records. WhatsApp messages, obtained under the Freedom of Information Act by Foxglove, show the firm pressuring Zubir Ahmed MP to help bring sensitive data into its platform (➚The Guardian).
Meanwhile, NHS England has deleted a claim from its website stating 137 hospital trusts are benefiting from Palantir’s AI health data system (The Guardian). The Guardian has previously revealed that 29 of the trusts were not using all or most of the technology. The original claim has now been replaced with a vague statement that users “already seeing real results — reduced admin, faster patient pathways, and more time for care”. We recently discussed the issue of overstating the performance of AI systems to justify their deployment in “With a Shovel of Salt: AI Success Stories and the Retraction Gap“.
Independent booksellers have said that unknown buyers are ordering unrelated titles in large quantities, prompting speculation about physical books being acquired in bulk for data extraction by AI companies. “It’s definitely a strange combination of books” one bookseller said, “Normally a book order is on a theme, but this is all over the place” (➚Tech Times).
A recent preprint paper asks what it means to use AI critically and examines how students critically evaluated AI-generated content. “The findings suggest that critical AI literacy extends beyond verifying the factual accuracy of AI-generated information.”
It finds that “critically engaging with AI involves situated and multidimensional judgment about the epistemic quality, contextual fit, practical feasibility, creative value, and ethical implications of AI-generated content.”
“By making the evaluative dimensions of the judgment more visible, this study offers a foundation for future research examining how critical AI literacy develops […]”.
H. Lee, & W. Yu (2026). What does it mean to use AI critically? Unpacking critical AI literacy through students’ evaluation of AI-generated content. arXiv:2610.10743
If you stumbled across this article on the web, subscribe to the “misaligned bits” newsletter and follow Misaligned on Medium, LinkedIn, Threads or Mastodon.
