What changed for your money this week.

Yes. Since Oct. 7, 2026, Claude Max includes $100 (Max 5x) or $200 (Max 20x) a month in Claude API credits, and Team plans get up to $500, per Anthropic’s own help centre. What the credit covers, why it skips Claude Code, how to claim it without linking the wrong organization, and what $100 buys. Verified Oct. 17, 2026.

Yes, it is real. On October 8, 2026 the FTC said it will send 2,124,796 PayPal payments worth more than $15.8 million to Cleo AI customers; the email comes from no-reply@consumersentinel.gov through October 26. Accept within 30 days. Our average: about $7.44.

Gemini is still free, but since Oct. 9, 2026 the free app sends most prompts to Flash-Lite, per Google’s own notice. What each plan gets now, why two Google pages disagree, and five $0 routes to AI Pro with the date each starts billing. Verified Oct. 10, 2026.

On September 17, 2026 the FBI warned that impersonators use AI to appear as police and officials on video calls; the scam drew nearly 61,000 complaints and more than $1.6 billion in losses from January 2025 to July 2026. We read all 11 FTC and FBI warnings from Sept. 1 to Oct. 6: 2 named AI.

ChatGPT and Claude both sell nonprofits a team seat for $8 a month against a $20 list price – 60% off, not the ‘up to 75%’ in the headlines. Who qualifies, the 20-person line and the premium-seat trap decide between them. Verified on each provider’s own page, Oct. 3, 2026.

YouTube’s ad tools find buyers for scammers the same way they do for honest advertisers. On Sept. 24, 2026 the FTC voted 2-0 to ask 132 questions about whether platforms must stop it. Google says it caught 99% of 8.3 billion bad ads in 2025; that still leaves up to 83 million that played first. What each ad button does.

OpenAI, Google, AWS and Zed are each giving academics free access to paid AI tools. The gates are completely different – one wants a preprint you authored, one takes PhD students but names no amount, one only counts about 130 listed universities. Verified on each provider’s own page, Sept. 28, 2026.

Paying for ChatGPT on the Billing page is an ordinary card subscription. Typing a card number into a chat is not safe. Letting an AI agent spend for you is unsettled: on Sept. 22, 2026 six banks published 16 recommendations that say should 28 times and must zero times. Until that changes, a credit card caps fraud losses at $50; debit can cost you everything.

US college and graduate students can claim four free months of ChatGPT Plus until Oct. 31, 2026. A card is required, it renews at $20 a month, and cancelling early forfeits the months you have not used yet – the opposite of most free trials. Verified on OpenAI’s own help centre, Sept. 19, 2026.

OpenAI and Anthropic both give verified US K-12 educators premium AI for nothing, and neither asks for a card. But the two free periods are built differently – one is a fixed date, the other a countdown that closes June 30, 2027 – and the June 2027 expiry still printed on page one of Google is a year out of date. Verified on the providers’ own pages, Sept. 14, 2026.

Yes, and federal rules allow it with conditions. On September 9, 2026 we read the enacted text of six state laws on AI in coverage denials: all six require a licensed human to make a medical-necessity denial, one (Utah) requires the insurer to tell enrollees it uses AI at all, and none require your denial letter to say whether AI touched your case. Arizona’s bill banned AI by name; the law the governor signed does not contain the word.

Yes, you can – and the drafting was never the legal part. On September 8, 2026 we read the will-execution statute in California, Florida, New York and Virginia: all four require two witnesses, and none says a word about who wrote the document. We also read the usage rules of four AI assistants (four of four require a licensed professional) and scored the 11 readable pages ranking for this question: three mention witnessing at all, and one names a number.

On January 14, 2026 the FTC finalized an order banning GM and OnStar from selling driver data to consumer reporting agencies. That order names three companies and leaves every other automaker to write its own rule. On September 7 we read six privacy notices from five other carmakers, and scored the eight readable pages that rank for this question: none of them mention the order, and none mention the free file you can demand to see whether your driving ever reached an insurer.

Amazon names three ways to get Alexa+, and the only plan called Alexa+ is the dearest of all of them. We priced every door from Amazon’s own plan pages on September 6, 2026: twelve months of the Alexa+ Standard plan is $239.88, twelve months of Prime — which includes the same Alexa+ plus free shipping and Prime Video – is $139.00. That is $100.88 more for less. Two doors cost $0, seven Echo models never get it at any price, and one state is written out of one feature by name.

Four days before the Sept. 9 event, the honest answer: nothing to buy, and five things standing in front of it. We read all ten footnotes on Apple’s Apple Intelligence page on September 5, 2026. The word ‘free’ appears zero times, a dollar sign zero times, and ‘subscription’ twice – both in footnotes, both attached to a paid iCloud+ plan. Plus the $400 gap between the two Apple Intelligence iPhones, and what Tim Cook actually said about iCloud.

A $19.99 line nobody remembers ordering, or $10 that appeared on a phone bill. We counted nine documented ways a Google AI Pro charge can start, from Google’s and Verizon’s own help pages on September 4, 2026 – and only five of them put anything on a Google statement. Plus the trap page one skips: accepting a free offer can cancel the plan you already pay for.

Between shoppers, generally no – the 1936 law everyone cites governs dealings between merchants. But on August 19, 2026 the FTC proposed an enforcement policy statement naming seven personalized-pricing scenarios as likely Section 5 violations, with comments open until September 18. We scored page one: four of six questions, zero answers.

Mostly no – and then, on January 1, 2026, yes for exactly 39 billing codes in six states. We counted the codes in CMS’s own operational guide and the people in CMS’s own enrollment file on September 2, 2026, and read the payment rule Congress went to the floor over.

The most common AI scams are deepfaked investment pitches, business email fraud, romance schemes, fake job interviews and the cloned-voice call from a relative in trouble – and in 2025 they cost Americans $893,346,472 across 22,364 complaints. But the ranking is inverted from the coverage: investment pitches took $632 million of that, the cloned-grandchild call took $5 million. A gap of 126 to 1. First edition of our monthly read of what the FBI and FTC actually published: the fake IC3 website the FBI warned about in July, the recovery scams that target people who already lost money, and our count of all eleven FTC consumer alerts from August 2026 – four of the eight scam alerts describe money that started moving after someone clicked a paid ad.

Canceling ChatGPT Plus takes four steps and about thirty seconds. Getting the $20 back is a separate screen with a separate answer, and for an American the honest answer is usually no. We read five AI subscription refund policies in full on August 31, 2026 – OpenAI, Anthropic, Google One, Google Play and Microsoft. Between them they name 12 countries, two trading blocs, and exactly one U.S. state: New York. Here are the exact cancellation steps for web, iPhone and Android, the 24-hour rule that decides your next bill, who to ask for a refund depending on who billed you, and why the federal click-to-cancel rule people describe to each other is currently a request for comments.

The plan that costs $69.99 instead of $99.99 still exists. It is just filed behind a button labelled Cancel subscription, and Microsoft has quietly deleted the only page that explained it. On August 30, 2026 we followed that page’s own URL: it 301-redirects twice into a generic article where the word “Classic” appears exactly twice, both times inside a proration table. Here is the route, the five conditions that close it, and the one setting that decides whether the option appears at all.

Look for the small white light near the right hinge: a brief blink is a photo, a light that keeps blinking is video. Meta has now patched that light three times since 2024, and the newest fix only started rolling out on August 28, 2026.

Google AI Pro now includes YouTube Premium Lite in the US at no extra cost – not full Premium, which is $15.99 separately. Verified US college students can get AI Pro and full Premium together for $8.99, the same price everyone else pays for Lite alone. Every figure read on Google’s own pages, Aug. 29, 2026.

We recorded all nine page-one Google results for the phrase people actually type. One belonged to Amazon; three were blog posts on a book-review site. Then we read six official contact pages: the federal ones print a number, the platforms print a chat box.

Four different systems get called “Meta AI using my photos,” and the guides on page one answer them as if they were one. On August 27, 2026 we read those guides, then read Meta’s own privacy pages and the request form they send Americans to — which turns out to cover third-party data, not your photos. Two of the four have a real off switch. One has a request reviewed “consistent with your local laws.” One has nothing. And the sentence that actually removes a photo from future training is sitting in Meta’s documentation, unquoted.

You look for the data-center line on your bill and there isn’t one, so you assume the story was hype. It isn’t — the charge is folded three levels down, inside a line called something else. PJM’s capacity price rose elevenfold in three years, and its own market monitor blames data centers for 38% of the latest $16.4 billion. We ran that through ComEd’s published rate formula, checked the result against two of ComEd’s own disclosures, and got roughly $5.56 a month — about $67 a year. Here is where it hides, the four clicks that show you your household’s own number, and the research that argues the opposite.

The advice everyone repeats — listen for the pause, the flat tone, the missing breath — was written against 2023 technology. On August 26, 2026 we read all seven guides Google puts on page one for this question, then read the voice vendors’ own specification sheets. Four of the seven still teach the listening test. The vendors publish that pause at 75 milliseconds, against a human conversational gap of about 200. Here is why the ear no longer settles it, why an AI call still does not have to announce itself under federal rules, and the four checks that do work.

OpenAI answers this on its own help pages: a deleted chat is scheduled for permanent deletion within 30 days — unless it has already been de-identified, or must be kept for legal reasons. On August 25 we read seven of those pages and found the sentence that undoes the tidy version: to remove something ChatGPT knows, you must “delete every source where it appears.” Here are all seven stores, which ones the Delete button actually reaches, the order you have to clear them in so memory does not refill from your history, and the block of 2025 conversations no setting of yours can touch.

The advice you remember — count the fingers, watch the lips — is now the worst advice. YouTube started labeling AI videos automatically in May 2026, so on August 25 we took 17 videos that are AI-generated footage by their own title and read the disclosure panel on every one. Seven carried the label. Three showed only an auto-dubbing notice that looks exactly like one. Seven showed nothing at all — including a 4.3-million-view compilation with “Sora Ai” in its title. Here is what the label actually says, why its absence proves nothing, and the four checks that still work.

Amazon’s own help page answers this one: “You can’t turn off these ads.” On August 24 we read the current Echo Show documentation, wrote down the menu path Amazon actually uses now — Display & Appearance, not Home & Clock — and checked every Alexa+ plan for an ad-free tier. There isn’t one, at $139 a year or $19.99 a month. Here are the six switches that genuinely shrink what rotates past you, what each one costs you in weather and reminders, and why the ads grew in the first place.