AI

  • |

    Meta Delays Its New Commercial Terms After Publishers Raise AI-Use Concerns

    ByLee Keybum

    Meta says it will keep consulting industry groups and business leaders before implementing proposed commercial terms that have drawn concern from publishers and digital advertisers. Reuters reported on October 28 that Meta gave no firm new implementation date. The proposed terms have not taken effect, so this is a delay in a dispute over wording—not evidence that Meta is already using content under the proposal.

    The calendar moved; the question did not

    The planned start date had already shifted from October 30 to November 20. Meta said consultation could continue beyond November 20, while existing terms remain in place in the meantime. That gives businesses more time before any change, but it does not settle what the proposed language means. In platform time, “we’re still talking” is both a pause and a reminder that the meeting invite belongs to the platform.

    There is no reported replacement date to circle on the calendar. The immediate change is procedural: Meta says it will continue consulting before moving ahead. For businesses trying to plan around platform rules, that leaves the central question open even as the deadline recedes.

    What publishers are worried about

    Publishers and digital advertisers have raised concerns that the terms could permit broad use, for AI services, of website content linked to business accounts. The News/Media Alliance has asked Meta to withdraw the proposed terms, and Nieman Lab has reported on the continuing disagreement over how the language could apply. Those are concerns about what the wording might allow—not confirmation that Meta has adopted or exercised those rights.

    That distinction matters because a proposed term, an interpretation of its possible reach, and a company’s actual use of content are not interchangeable facts. The current story is about the wording and what different parties think it could mean, not a finding that Meta has begun using business-linked website material in the way critics fear.

    Meta disputes that reading

    Meta rejects the interpretation that its proposed terms would grant the broad content-use permission critics fear. That disagreement matters: a company’s explanation and a customer’s reading of the terms can both be part of the story, but neither turns the disputed interpretation into settled policy. For now, Meta says it is consulting and the existing terms stay in force. The question of what content could be covered if the proposal moves ahead remains unresolved to the satisfaction of critics.

    The fine print has a power imbalance

    A business that relies on a large platform may have little appetite for treating every terms update like a weekend seminar in contract interpretation. The platform controls the proposed wording; the business has to work out what it means for its online presence and the content connected to its account. That is the Terms of Surrender feeling—not because the disputed permission is already in use, but because businesses are being asked to plan around language that is still being argued over.

    The useful takeaway is modest: the rollout is delayed, existing terms remain in place, and the disagreement is unresolved. More consultation is better than a surprise switch, but a calendar extension is not clarity. The next meaningful step will be whether the final language makes its boundaries understandable to the businesses expected to live with it.

  • |

    AI’s Public Stakeholders Are Still Waiting on the Check

    ByLee Keybum

    AI’s stakeholder agreement has familiar fine print: the public helps lay foundations, then gets treated like a guest when the upside is discussed. Public research, grants, labs and universities can help make future innovation possible; that doesn’t mean any one grant built any one company’s product. It does mean the private-first habit deserves scrutiny. Public contribution should count when AI’s benefits are shared, not fade into the background while wealth and control stay private.

    A public dividend is a proposal, not a check already in the mail. But if taxpayers are important enough to appear in AI’s origin story, they’re important enough to count in the conversation about who benefits. Apparently we got a seat at the launch and missed the dividend meeting because the calendar invite went to founders only.

  • |

    The Public Helped Take AI’s Early Risk. It Deserves a Dividend.

    ByLee Keybum

    When public grants, labs, universities, or taxpayer-funded research help shoulder AI’s early risk, it’s odd to treat the upside as an investors-only boarding pass. I read the terms: innovation deserves a reward, but the public shouldn’t be the silent partner that helped build the runway and got left at the gate.

    An American dividend is a proposal, not an existing policy or settled funding plan. Still, if the public helped pay for takeoff, it deserves more than a thank-you and an in-flight safety card while private investors get the view from first class. Big Tech can call that innovation; ordinary people might ask who picked up the tab—and why they weren’t invited to share the return.

  • |

    OpenAI Built a Time Machine for Job Applications

    ByLee Keybum

    Lee Keybum has read enough terms of service to know that the future usually arrives with a checkbox, a data harvest, and a subscription barnacle under the floor mat. OpenAI sells an automated tomorrow, but the Justice Department says some U.S. applicants for certain PERM positions allegedly had to navigate a paper-based obstacle course to apply for jobs there. The DOJ said those positions were not listed on OpenAI’s public job site, required paper applications, and were promoted in ways that discouraged U.S. workers. The company building tools to remove administrative friction had apparently placed a toll booth in front of its own hiring process.

    That is the corporate version of inventing a robot to carry groceries, then making the customer drag the bags home because the robot is reserved for management. The point is not that every OpenAI job used this process, or that the settlement proves every allegation beyond dispute. The point is the spectacular mismatch between the product pitch and the alleged user experience: artificial intelligence for the world, paper archaeology for the applicant.

    On August 4, the DOJ announced a $3.2 million settlement finalized the day before. Its terms include a $1.2 million civil penalty and a $2 million back-pay fund, along with electronic applications, public job postings, employee training, and monitoring. In other words, the paperwork eventually generated enough paperwork to require a second, more modern paperwork system.

    OpenAI did not need a time machine to reach 1998. It allegedly just needed a hiring department that treated “please find the hidden opening, print the form, and hope someone receives it” as an acceptable interface. Meanwhile, the rest of the company is helping businesses automate scheduling, sorting, drafting, and the other chores ordinary workers have been told will be transformed by software. The cloud owns cab fare, but apparently applicants still had to walk to the office.

    Here is the practical audit: if automation is advanced enough to reorganize everybody else’s work, it should be advanced enough to let a qualified person find the door and apply electronically. The DOJ settlement does not prove a broader corporate philosophy, but it does expose a familiar one: friction disappears fastest when it costs the company money. For everyone else, the login ate the afternoon, the posting moved into a drawer, and the future arrived wearing a fax-machine costume.

  • |

    ChatGPT Wants Your Medical Records, But Not the Responsibility

    ByLee Keybum

    Lee Keybum reads the terms so you do not have to, and OpenAI’s new Health feature comes with a very modern trust bargain: ChatGPT can become familiar with your connected medical records, lab results, prescriptions, Apple Health data, and health conversations, but it still wants to be treated as support rather than medical care. OpenAI announced the U.S. rollout on July 23, describing Health in ChatGPT as a place for personalized guidance and health-related conversations. That is a lot of intimate information to hand to a general-purpose chatbot before it politely reminds you that it is not intended to diagnose or treat anything.

    The contradiction is not hidden; it is laminated into the product. The more information ChatGPT can use, the more naturally it can sound like the friend who knows why every pill is in the cabinet and which lab result made you stare at the ceiling. But OpenAI says the service does not replace qualified medical professionals. In ordinary-person language, the platform wants the context required to sound medically informed while preserving an escape hatch if the conversation wanders into actual medical responsibility.

    OpenAI’s health privacy policy also says connected health data is not used to train its foundation models or target ads by default. That matters, and “by default” matters too. It is not a promise that every possible privacy concern has evaporated into the cloud. It means the company’s stated bargain is narrower: users may authorize sensitive data connections for the feature, while OpenAI says those connections are treated differently from ordinary ChatGPT data for model training and advertising.

    For users, the practical question is not whether the chatbot can produce a soothing paragraph about wellness. It is whether convenience quietly turns the app into the most informed entity in a person’s medical life without giving that entity the accountability people normally expect from medical care. A doctor has credentials, professional rules, and a human being attached to the decision. ChatGPT has a privacy policy, a disclaimer, and a remarkable talent for making a sentence sound settled before the facts are.

    That is the subscription-barnacle version of artificial intelligence: first it asks for your bloodwork, sleep history, medication list, and trust; then, if the answer is wrong, it becomes a very confident autocomplete with no medical license. OpenAI may call Health a support tool, but the user experience is built to feel personal and informed. The company wants the data that makes the system sound like a doctor while keeping the responsibility label safely out of frame.

  • |

    The White House Has Entered Its AI Cartoon Villain Era

    ByHolden McGroin

    My corkboard has one new pin: the White House reportedly posting an AI-generated, Rick and Morty-style political clip featuring Donald Trump, J.D. Vance, Tom Homan, and Chuck Schumer before ending in a glowing MAGA flourish. Euronews reported the clip sparked major backlash, which is a polite way of saying the presidency wandered into the internet wearing a fake mustache and asked the algorithm for validation. This is not merely cursed cartoon sludge. It is official communication borrowing the machinery of viral culture-war bait.

    The formula is simple: recognizable figures, synthetic spectacle, tribal signaling, and just enough outrage to make everyone else distribute the message for free. The Daily Beast separately reported on a video in which Trump turns a political nemesis into a robot, suggesting the administration’s preferred media department may now be staffed by a teenager with a render farm and an unresolved comment-section grievance. The point is not that a cartoon clip replaces every serious government function. The point is that public messaging is being optimized for reaction before it is optimized for explanation.

    That is the contradiction sitting in the middle of this whole production. A presidential institution is supposed to communicate about public business that affects workers, families, patients, students, and people who spend half their lives on hold with a government office. Instead, the feed gets a Rick and Morty-style AI sequence with Trump, Vance, Homan, Schumer, and a glowing MAGA finale. The administration may call this internet fluency. Ordinary citizens are allowed to call it a strange use of the national attention span.

    Follow the thread but check the knot: who benefits when government communication becomes a permanent outrage dispenser? The platform gets engagement, the faction gets a loyalty ritual, and the officials get to look busy without explaining the boring machinery that determines whether life gets easier or more expensive. Everyone else gets dragged into the group chat to referee a synthetic cartoon while the actual public business waits outside with a clipboard.

    The White House wanted to look like it owned the internet. Instead, it made the presidency look like a federal-budget group chat run by people chasing engagement. That is the real panic story: not that AI can make a cartoon villain, but that public power is learning to speak like one.

  • |

    Google Wants AI Speed With Human Job Security

    ByLee Keybum

    Google workers are asking for job-security protections while the company asks them to build the future, which is a little like being hired to construct a lifeboat and then receiving a calendar invite titled “Headcount Efficiency.” KQED reports that roughly 4,500 Google employees signed a petition seeking protections including guaranteed severance and extended paid-leave options. The petition was delivered at a Mountain View rally on July 16, turning corporate risk into something with shoes, rent, and a household emergency fund.

    Management language often presents layoffs as difficult decisions, as if employment were weather and executives were merely standing outside with a clipboard. But a layoff is not fog. It is a decision that moves the company’s uncertainty onto the worker’s kitchen table. The corporation keeps the flexibility; the employee gets to discover whether the login ate his afternoon, his health coverage, or next month’s grocery budget.

    That is the contradiction Google’s AI-era confidence cannot calculate away. The company depends on people to build, maintain, test, and support products increasingly tied to artificial intelligence, while workers say continued employment is treated as an unreasonable promise. KQED’s report notes that employees pointed to Google’s rising valuation and profitability as part of their criticism. The basic question is not whether a company may ever reorganize. It is why the people doing the work must absorb the entire shock whenever management reorganizes.

    The Communications Workers of America also used the latest Alphabet and Meta layoffs to call on tech workers to organize, while another report from PC Gamer described the petition as a protest against profit being placed above the people who make the company run. Those are not demands for a crystal ball or a lifetime appointment in Mountain View. They are requests for a softer landing when a highly profitable corporation decides the human beings attached to its products have become inconvenient numbers.

    Google can forecast machine intelligence, sell confidence about tomorrow, and make the cloud sound like it owns cab fare. Yet its most advanced prediction system still appears unable to answer the oldest workplace question: “Will I still work here on Monday?” That is not futuristic management. That is a subscription barnacle wearing a headset and calling the cancellation policy innovation.

  • |

    Reflecting Pool Dipper: AI-enhanced “suspect hunt” meets reality (and loses)

    ByHolden McGroin

    Meet my newest house pet: the Suspect Hunt Goblin. It only gets excited when someone says, “Don’t worry—enhance it. You can tell it’s obvious now.” Then it scampers straight into group chats like: “Guys, it’s clearer, so it counts. This is basically sworn testimony with better lighting.”

    Here’s the contradiction audit, straight from the storyline: US Park Police shared distant footage for a “Destruction of Government Property” investigation, and the online vibe-check sprint decided that distance + uncertainty could be converted into a name once the image got AI-polished. But Lead Stories reportedly pushed back that the viral AI-enhanced stills weren’t reliable for identifying the person in question. In other words: the thing people used to claim certainty wasn’t actually good enough for the identification they wanted it to do.

    So who benefits from the “enhance-and-apprehend” loop? Not truth. Not verification. The benefits mostly go to the feeling machine: armchair detectives get to feel involved, the outrage engine gets momentum, and everyone gets to cosplay as reality’s detective—without doing the hard part, which is accepting that “unclear” stays unclear, no matter how many filters join the chat.

    The panic doesn’t reduce uncertainty. It just upgrades it into digital certainty cosplay—and then everyone pretends that’s the same as evidence.

  • |

    The Privacy Settings Keep Getting Smarter Than the Users

    ByLee Keybum

    The newest trick in tech is to make privacy sound like a premium feature, which is a bold move for something users thought was included when they said yes to the app. One day it’s an AI helper; the next, it’s a subscription, a policy update, and a little lecture about “improving your experience,” which is corporate for “please enjoy the machine learning while it learns you.”

    That’s the modern deal: companies promise convenience, then quietly reclassify your habits as an asset class. The user gets a smarter feed, a pricier plan, and a privacy page long enough to qualify as light reading for a tax attorney. If that’s innovation, it’s at least honest about the new product: you, but organized for monetization. Share with someone who still thinks “free” is a setting, not a prequel.

  • |

    Meta Data Labelers Unionize, Then Lose 1,100 Jobs—Was AI or Union Angst to Blame?

    ByLee Keybum

    They raised a union flag—then the boss yanked it and the floor got cleared. In early April 2026, data-labeling contractors at the Nairobi office of Sama, working for Meta, voted to unionize, aiming to address issues like low pay, mental-health strain, and job instability. Fast forward a few weeks, and Meta terminated its contract with Sama, resulting in approximately 1,108 to 1,110 workers facing layoffs. Meta cited ‘automation’ and ‘shifting project needs’ as the reasons. Sounds a bit like the platform toll booth just snatched another round of rent.

    Why should you care? These contractors weren’t just sorting any old data—they were putting Meta’s Ray-Ban smart-glasses through their paces. You know, those nifty glasses that were supposed to be all about discreet video recording? Turns out, the workers at Sama raised concerns after seeing footage that wasn’t exactly family-friendly viewing material—think privacy-invading moments recorded without people’s knowledge, in, shall we say, quite intimate settings.

    The whistleblowers claimed they stumbled upon videos of people in bathrooms, undressing, and, yes, engaging in activities best left to the imagination. Meta’s promise of private recordings just got as private as your lunch table at a food court. The seriousness of these claims isn’t lost, as both U.K. and Kenya regulators have launched investigations, and there’s even a class-action lawsuit in California circling around the glasses’ prying ways.

    Meta’s official stance is about as surprising as a Terms of Service update: they blame the layoffs on project needs and automation, with a side of ‘standards not met’. Sama, on their part, denies dropping any balls, much like a juggler at a tech-themed circus. But the timing here is as questionable as the juice cleanse diet industry.

    The fallout is stretching its legs beyond Meta’s walls. The legal and regulatory spotlight is beaming down on this mess, shining through any carefully curated corporate message like a laser through fog. And while Meta’s firing comes with all the unconvincing necessity of the obligatory “Agree” button click, one can’t help but notice how quickly union voice can be silenced when automation claims are waved around like a magic wand.

    So, the next time you toss on your Ray-Bans, remember: tech isn’t just lines of code; it’s propped up by a web of global labor that’s often as visible as your Wi-Fi signal. Human oversight is involved—sometimes unequipped, sometimes unnoticed, but always at risk of being swept away by the algorithmic shrug.

    In an industry where your data might get more privacy than those gathered to sort it, it’s worth a laugh, albeit through gritted teeth. If the login keeps eating your afternoon, at least make the Terms of Surrender worth your time.

    Sources

End of content

End of content