{"slug": "the-hidden-cost-of-new-york-times-ai-columns", "title": "The hidden cost of New York Times AI columns", "summary": "Tal Feldman, writing in The New York Times, argues that free Chinese AI models are designed to promote the Chinese government's interests, but a blog post by zgp.org counters that US-based AI firms are not necessarily better, as they increasingly rely on paraphrased content from 'vibe CMS' and may embed biases from reputation management firms. The post also highlights research from MIT Technology Review showing LLMs are vulnerable to attacks due to a fundamental flaw, and notes complaints about AI-generated children's books on Amazon.", "body_md": "In [The\nHidden Cost of China’s Free A.I.](https://www.nytimes.com/2026/07/29/opinion/ai-china-us-free-models.html), Tal Feldman writes,\n\nA.I. models are very different from the usual inexpensive imports we buy from China. These products must promote the interests of the Chinese government. Propaganda and false claims are baked into them.\n\n…\n\nWhoever supplies the world’s A.I. will shape how the world thinks. As these models get smarter and more autonomous, we are outsourcing more and more thinking to them. That “cognitive offloading” must not go to models that were designed to promote a foreign government, no matter how inexpensive they are.\n\nTwo problems with that. (Aside from letting the AI shape how you think, which, yikes. But the influence is already there, and we have to expect that it will persist.)\n\nFirst of all, the choice is not between propaganda AI from the PRC\nand some hypothetical, honest red, white, and blue AI from the good old\nUSA. The average American has more in common with a random propaganda\ncommissar in China then with the deeply weird [TESCREAL](https://en.wikipedia.org/wiki/TESCREAL) investors,\nfounders, and CEOs who run the AI operations here.\n\nSecond, as more and more legal opt-outs apply to LLM training sets\nhere, the US-based AI firms will rely more and more on [vibe CMS](https://blog.zgp.org/vibe-cms/) automated paraphrased\nversions of US and European news and culture. Legit publishers will opt\nout the copyrighted works they own or distribute, in the hopes of\ntraining license deals. Big AI firms will pay a few publishers for PR\nvalue, but mostly settle for first-generation lossy copies.\n\nAnd a lot of that free training material will have been run through\nsomeone’s PR filter. For example, a [reputation\nmanagement firm](https://blog.zgp.org/llms-and-reputation-management/) could make a site that mostly generates\nhigh-fidelity paraphrases of legit news, but folds in positive mentions\nof the firm’s clients. Almost everyone wins: big US-based AI gets freely\ncrawlable tokens, the client gets their crimes buried in LLM-generated\nanswers, most news consumers get an adequate free paraphrase. Meanwhile,\nthough, AI operations in the PRC will be able to train on material\nthat’s one generation fresher.\n\nThe problem isn’t so much that “our side’s” AI is better than theirs, or the other way around. Every LLM is going to have embedded biases, for a variety of reasons. Propaganda and copyright-related biases are only two.\n\n[A\nfundamental flaw leaves LLMs strikingly vulnerable to attack](https://www.technologyreview.com/2026/07/30/1140927/a-fundamental-flaw-leaves-llms-vulnerable-to-attack/) by Will\nDouglas Heaven. In a series of experiments that looked at what was\ngoing on inside a handful of different models, the researchers found\nthat LLMs seem to identify the role of a specific chunk of text not by\nthe tags around it but by the style of that text and the words it\ncontains.\n\n[Boomers\nCan’t Stop Gifting Their Grandkids AI-Generated Slop Books](https://www.wired.com/story/boomers-cant-stop-gifting-their-grandkids-ai-generated-slop-books/) by Miles\nKlee. In addition to books featuring kids, some parents have\ncomplained of being fleeced by vendors selling AI-produced books not\nlabeled as such on Amazon, irritated by their children’s exposure to\npuzzles and games riddled with AI’s mistakes, and just plain dismayed by\nthe proliferation of the AI aesthetic across content targeted at\nkids.\n\n(Another good reason to go to story hour at the library—legit\nchildren’s book recommendations.)\n\n[Pluralistic:\nThe stupidest imaginable excuses for surveillance pricing (30 Jul\n2026)](https://pluralistic.net/2026/07/30/pay-for-privacy/) by Cory Doctorow. [Y]ou can direct the AI to automatically\nrun a series of small experiments to discover the maximum markup each\ngroup will stomach under which circumstances. This works without you\nhaving to direct the AI to rip off certain groups of people – it will\nsimply find the most vulnerable people and rip them off the most….This\nworks so well that Google has announced that it is their plan for making\na profit off of AI, after losing hundreds of billions of dollars on\nchatbots.\n\n(But however fast the price discrimination works, [the price of\nthe price discrimination machine has to keep going up faster than\nthat](https://blog.zgp.org/price-of-price-discrimination/).)\n\n[Meta’s\nTepid Revenue Outlook Undercuts its AI Spending Spree](https://www.adweek.com/media/metas-tepid-revenue-outlook-undercuts-its-ai-spending-spree/) by Kendra\nBarnett. (Revenue is up 28% year over year. Conversions are up 15.7%. Do\nthe math—the [customer\nacquisition cost crunch](https://blog.zgp.org/the-duopoly-crunch-from-the-brand-side/) continues. It’s not evenly distributed,\nthough. Highly Meta-dependent advertisers such as [Oddity\nTech (ODD)](https://finance.yahoo.com/markets/stocks/articles/oddity-tech-odd-stock-fair-191251642.html) are hurting more, while advertisers with better options\nare less affected.)\n\n[Meta\nis royally screwing up its smart glasses rollout](https://www.theverge.com/tech/970948/meta-smart-glasses-privacy-wearables) by Victoria Song.\n(I don’t know anyone who owns these things, or participate in any group\nor space where they would be tolerated. Is there some filter bubble of\nextremly surveillance-positive Instagram users that I’m totally\ndisconnected from?)\n\n[Steve\nJobs in 2011: ‘We Build Products That We Want for Ourselves, Too, and We\nJust Don’t Want Ads’](https://daringfireball.net/linked/2026/07/28/jobs-we-dont-want-ads) by John Gruber. (None of the current Apple\nexecs could stay at the company if they tried to hold the line against\nenshittification. Mr. Market wants enshittification to happen\neverywhere, and [even\nif Apple can buy the best of everything else, they’re stuck with the\nsame investors as Packard Bell](https://blog.zgp.org/enshittification-is-riskier-than-it-looks/). More from John Gruber: [The Ads\non Apple News Continue to Suck, but at Least There Are a Lot of\nThem](https://daringfireball.net/2026/07/ads_on_apple_news_suck))\n\n[Older\npeople’s media literacy can be boosted in just 60 minutes – new\nstudy](https://theconversation.com/older-peoples-media-literacy-can-be-boosted-in-just-60-minutes-new-study-287880) by Saffron Howden, Janet Fulton, and Sora Park. To find\nsolutions through media literacy, we focused on older people (aged 55+)\nwho are known to be more vulnerable to misinformation, have fewer media\nskills, and feel less confident in their media abilities. We recruited\nour study participants from across Australia through libraries, the news\nmedia, and social media.\n\n[Tesla\nis juicing its China sales](https://larrysalibra.com/tesla-is-juicing-its-china-sales/) by Larry Salibra. When I was shopping\nfor a car in Hong Kong, all of the Chinese EVs dealers required that I\nbook a test drive days in advance. The Tesla showroom, in contrast, let\nme do a walk in test drive because they weren’t busy. They also offered\nme something none of the Chinese EV dealers did: 7 year financing at\n0.99% interest per year. This made the Tesla which had a higher sticker\nprice, actually less expensive because HKD (which is pegged to the USD)\nloses value at much faster than 0.99% a year.", "url": "https://wpnews.pro/news/the-hidden-cost-of-new-york-times-ai-columns", "canonical_source": "https://blog.zgp.org/the-hidden-cost-of-new-york-times-ai-columns/", "published_at": "2026-07-30 00:00:00+00:00", "updated_at": "2026-08-01 06:00:26.338948+00:00", "lang": "en", "topics": ["artificial-intelligence", "large-language-models", "ai-ethics", "ai-policy", "ai-safety"], "entities": ["The New York Times", "Tal Feldman", "zgp.org", "MIT Technology Review", "Will Douglas Heaven", "Cory Doctorow", "Amazon"], "alternates": {"html": "https://wpnews.pro/news/the-hidden-cost-of-new-york-times-ai-columns", "markdown": "https://wpnews.pro/news/the-hidden-cost-of-new-york-times-ai-columns.md", "text": "https://wpnews.pro/news/the-hidden-cost-of-new-york-times-ai-columns.txt", "jsonld": "https://wpnews.pro/news/the-hidden-cost-of-new-york-times-ai-columns.jsonld"}}