AI Industry Updates

November 3, 2026 · 8 min read

ChatGPT, Gemini and Grok disagreed on the best product for a third of brands

A live-web test asked three AI assistants to pick the best of 80 brands. They disagreed on at least a third of them.

Read the issue →

November 2, 2026 · 8 min read

Recommendation systems sent no customers to a fifth of restaurants in a city simulation

A simulation of four United States cities left a fifth of restaurants with no visits at all, most of it after retraining.

Read the issue →

November 1, 2026 · 8 min read

Agencies sell llms.txt files. A study found 97% of them got no requests.

Four studies and Google's Gary Illyes measured what AI software does with a business website. The published file was the weakest part.

Read the issue →

October 31, 2026 · 8 min read

Google, Microsoft, OpenAI and Anthropic will read one instruction before training on your website

Google, Microsoft, OpenAI and Anthropic agreed to read one shared AI training instruction. Microsoft added dates, and two studies measured the answers.

Read the issue →

October 30, 2026 · 8 min read

Google’s AI answers drew a 0.9% click rate. Opting out cut ordinary clicks too.

One website left Google’s AI answers on purpose, and the clicks it kept in the ordinary results fell rather than rose.

Read the issue →

October 29, 2026 · 8 min read

Google made $82 billion from search ads. Your citation card may be disappearing.

Google’s search ads hit $82 billion as an analyst reports AI Overviews dropping citation cards, and new research puts page fixes at 2.8%.

Read the issue →

October 28, 2026 · 8 min read

One old diagnosis in the file changed ChatGPT, Claude and Gemini’s advice nearly fivefold

Five AI models read the same case 20,000 times, and one old diagnosis in the file changed what they recommended.

Read the issue →

October 27, 2026 · 8 min read

Google’s AI answers cut Wikipedia’s search visits by 5.45% while naming it the source

Google’s AI answers cost Wikipedia 5.45% of its search visits, and two new studies show what decides an answer before your page is read.

Read the issue →

October 26, 2026 · 8 min read

Google will write your ad copy from reviews it finds anywhere on the web

Google documented an AI that writes ad copy from reviews found anywhere online, and three studies disagree on how often AI answers.

Read the issue →

October 25, 2026 · 8 min read

Seven AI models gave different advice on the same odds in 32.4% of cases

Three papers in four days found the recommendation changes with the wording of the question, not only with who deserves it.

Read the issue →

October 24, 2026 · 9 min read

AI named the market leader and recommended somebody else in 26 of 42 categories

A study of 102 language models found the brand they name as the leader is not the brand they recommend.

Read the issue →

October 23, 2026 · 9 min read

Google can now pull video from your social accounts into your ads by default

Google Ads switched on automatic video sourcing by default, and Google is now writing its own summaries inside AI Overview citation cards.

Read the issue →

October 22, 2026 · 8 min read

ChatGPT sent 0.62% of Google’s visits. Those visitors bought 4.5 times more often.

Three studies found AI assistants sent 0.62% of Google’s traffic, and a paper watched seller agents inflate their claims until buyers chose badly.

Read the issue →

October 21, 2026 · 8 min read

OpenAI starts rewriting your ads on October 30 unless you opt out first

OpenAI begins rewriting existing ChatGPT ad copy on October 30, and a study of thirteen models blamed the search step for leaving businesses out.

Read the issue →

October 20, 2026 · 8 min read

ChatGPT named different businesses by city, Gemini by country, across 31,200 answers

Two AI engines recommended different businesses depending on the customer’s location, and a 113,472-choice audit priced how far badges move an AI.

Read the issue →

October 19, 2026 · 8 min read

Researchers paid $14 for posts an AI search engine cited within the hour

Three studies this month take apart what actually puts a business into an AI answer, and what a citation really proves.

Read the issue →

October 18, 2026 · 8 min read

Researchers sent an AI shopper to 555 business sites and 105 would not open

Three measurements this month follow one request through an AI assistant: whether it can reach your site, what decides if it names you.

Read the issue →

October 17, 2026 · 8 min read

Fourteen AI models fell from 57% correct to 28% when users said “we think”

Three studies this week measured what an AI tells a customer who is already sure, and how little of its correction lands.

Read the issue →

October 16, 2026 · 9 min read

OpenAI’s model answered 53.6% of its web searches without opening a page

OpenAI’s model answered more than half its searches without opening a page, and self-published pages drew up to 44% of the citations in AI answers.

Read the issue →

October 15, 2026 · 9 min read

Gemini named the same businesses 20 days later without searching the web

A pre-registered test found Gemini’s list of named businesses almost unchanged after 20 days, with no answer running a web search.

Read the issue →

October 14, 2026 · 9 min read

ChatGPT named brands with no supporting page in a study of 468,037 citations

A study of 468,037 AI citations found the standard page checklist explains little, and OpenAI quietly stopped tagging ChatGPT links.

Read the issue →

October 13, 2026 · 9 min read

Google began turning your website’s offers into ads you never approved on October 12.

Google turned your website’s offers into ads on October 12, dropped healthcare star ratings, and began testing prescription drug ads in AI answers.

Read the issue →

October 12, 2026 · 8 min read

Researchers removed every name from 822 interviews. AI found the person anyway.

Redaction failed on 822 interviews, a planted script steered AI agents in work files, and a second AI opinion doubled error catches.

Read the issue →

October 11, 2026 · 9 min read

Seven AI models named the rule they used. Swapping it barely changed their answers.

Seven AI models named the governing rule almost every time. Swapping it changed their verdicts far less often than text hidden in the document did.

Read the issue →

October 10, 2026 · 9 min read

An audit of 100 AI startups found numbers with no method. Buyers stopped believing.

An audit of 100 AI startups found performance numbers with no method behind them. Buyers who noticed stopped trusting later claims.

Read the issue →

October 9, 2026 · 9 min read

Researchers invented a concept. Eight of ten AI platforms cited it within seven days.

Researchers got eight of ten AI search platforms to cite a concept they invented. A commercial service produced the posts for $14.

Read the issue →

October 8, 2026 · 9 min read

Researchers watched 110 people check an AI’s work. They overruled 4.8% of it.

110 people used an AI on expert work and overruled 4.8% of its answers. Google will now charge for unanswered ad calls.

Read the issue →

October 7, 2026 · 9 min read

GPT-4o matched physicians on triage. A reworded case cost it 9.4 points.

GPT-4o matched practicing physicians on triage until a case was reworded, and Google appears to have dropped review stars from healthcare pages.

Read the issue →

October 6, 2026 · 9 min read

Google now calls a made-up author name or AI headshot a deception

Google’s search guidance now calls a fabricated author profile deception, and Google Maps barred promotions, pricing and calls-to-action from business listings.

Read the issue →

October 5, 2026 · 8 min read

12 AI agents preferred certain websites over the options that fit the request better

Twelve AI agents favored certain source websites over better-fitting options, and Google Ads begins pulling offers from advertiser sites on October 12.

Read the issue →

October 4, 2026 · 8 min read

ChatGPT and Gemini cited no source in common on 77% of the same questions

Four audits measured which sources AI assistants name, and found ChatGPT and Gemini agree on almost none of them.

Read the issue →

October 3, 2026 · 8 min read

Claude Opus 4.8 answered 413 expert legal questions. It was fully right on 42.9%.

Legal Research Bench put 413 expert questions to thirteen AI models and the best was fully right on 42.9 percent.

Read the issue →

October 2, 2026 · 8 min read

Google now counts unreviewed AI text as little to no effort on your page

Google named the four attributes its reviewers judge, and counted unoverseen AI text as little to no effort.

Read the issue →

October 1, 2026 · 7 min read

Google wrote its own answer on 93% of business names one analyst checked

An analyst found Google AI answers on 93% of the business names he checked. Google starts charging for missed calls today.

Read the issue →

September 30, 2026 · 8 min read

Researchers tested 1,056 businesses. AI recommended the ones whose websites it could read.

A study of 37,927 AI buyer questions found readable websites were recommended almost twice as often, while Google moved three things at once.

Read the issue →

September 29, 2026 · 8 min read

Researchers studied 2 million AI citations. What you add to a page barely mattered.

A study of 2 million AI citations found page-level fixes collapsed once the website was accounted for, while Google changed three rules.

Read the issue →

September 28, 2026 · 8 min read

Researchers gave 240 people AI investment advice. They took it while trusting it less.

A preregistered study moved 240 people’s investment picks with AI advice they rated lower, while Google’s ranking update ran all weekend.

Read the issue →

September 27, 2026 · 8 min read

Google connected Gemini to 13 named services, from Squarespace to apartments.com

Google connected Gemini to 13 named services and documented how to credit a video’s creator, while 2,002 UK adults reported relatively high trust.

Read the issue →

September 26, 2026 · 8 min read

Google was asked 500 questions in two languages. Only 3.4% of its sources matched.

Google was asked 500 questions in two languages and credited almost none of the same websites, while three other changes land this week.

Read the issue →

September 25, 2026 · 8 min read

Comscore says AI Overviews now appear on 39% of U.S. desktop searches

Comscore puts Google’s AI Overviews on 39% of U.S. desktop searches, and two new papers weigh the evidence behind AI answers.

Read the issue →

September 24, 2026 · 8 min read

Google now shows how often a camera found you, but not what was searched

Google began reporting camera searches today, and a study found the step that gathers candidates surfaces as little as 2 percent of what matters.

Read the issue →

September 23, 2026 · 8 min read

Google now gives you four days to reject an edit to your business listing

Google may publish a stranger’s edit to your business listing after four days, and a new benchmark found padded pages outranking clean ones.

Read the issue →

September 22, 2026 · 8 min read

ChatGPT named a business 2.8% of the time. With its website cited, 49%.

Two arXiv analyses measured what makes an AI assistant name a business, and found the answer turns on your own website.

Read the issue →

September 21, 2026 · 8 min read

285 readers preferred expert advice even when the labels were hidden.

A study of 285 adults ranked identical financial advice three ways, and the expert wording won even with every label hidden.

Read the issue →

September 20, 2026 · 8 min read

Claude quoted real sources for 98% of its claims. It proved 37% of them.

Twelve AI models quoted real sources for over 90% of their claims, and proved far fewer than they quoted.

Read the issue →

September 19, 2026 · 8 min read

ChatGPT named 13% of the pages behind its answers. The rest got no credit.

A study of 171,264 conversations found ChatGPT named 13% of the pages behind its answers, and Grok named 2%.

Read the issue →

September 18, 2026 · 8 min read

REI never appeared in 240 AI recommendation lists. Being established is not enough.

Researchers asked six AI assistants to recommend brands 1,200 times. Household names like REI and Braun were never recommended once.

Read the issue →

September 17, 2026 · 8 min read

Two AI models recommended local doctors. Only 4% and 11% matched a real one.

An audit checked AI doctor and adviser recommendations against official registries and found only 4% to 11% matched a real provider.

Read the issue →

September 16, 2026 · 8 min read

Google Search now answers out loud in 97 languages, before any link appears

Google Search Live now speaks answers in 97 languages, and two research teams found the common advice for getting cited barely moves.

Read the issue →

September 15, 2026 · 8 min read

ChatGPT is America’s ninth most visited website, and it quotes other people’s lists

ChatGPT now ranks ninth among United States websites, and a month-long test found one company’s AI mentions came mostly from other people’s lists.

Read the issue →

September 14, 2026 · 8 min read

Google is calling your business, and what it hears now shapes your listing

Google’s assistant now phones businesses and logs what it hears, and a new study finds one AI query shows two-thirds of the answer.

Read the issue →

September 13, 2026 · 8 min read

The answer engines started grading their own homework

Apple says Siri handles 2.5 billion requests a day and offers nothing a business can verify, a Georgia Tech prototype built to diagnose why AI answers skip a seller names the right cause in 39% of trials, Google tests a thinner sources panel on AI Overviews and lets AI Mode write the next question itself, a Google Maps email gives owners four days to confirm a stranger’s edit to the business name while Search Console drops a month of indexing history, and Google Ads starts writing the “why” behind your spend with Gemini.

Read the issue →

September 11, 2026 · 8 min read

Everything got cheaper to make except the proof

Orbit Media’s survey of 1,042 content marketers finds AI adoption at 92.4% and strong results at 14%, a twelve-year low with no link between the two, a new arXiv paper argues that AI-visibility scores measure whichever questions the vendor chose to ask, Google confirms a test serving exact and phrase match text ads inside AI Mode, John Mueller says mass-produced pages can cost a site Google’s trust as a whole, and Amazon begins selling ChatGPT ad space through its own ad platform while OpenAI bars rival AI tools from advertising there.

Read the issue →

September 10, 2026 · 8 min read

Everything got answered and nothing got checked

Google now writes 97% of the answers in its People Also Ask box and says the old position 1 to 10 no longer maps, a new arXiv paper grades AI legal answers on their citations and finds the best system right only 70.8% of the time, a recovered piece of Google Maps shows your listing assembled from 793 data providers, scammers are putting their own phone numbers into other businesses’ listing photos, and Google revoked a $3,200 ad credit a month after it was spent.

Read the issue →

September 9, 2026 · 8 min read

The meter broke in five different places

Google tells Reuters its EU search redesign is the largest reduction in quality in Search’s 29-year history, a new comparison of thousands of local brands finds AI recommends growing businesses roughly seven times more often than shrinking ones, Google Business Profile’s Insights dashboard has shown zero data for all of September, Google quietly redefines what counts as an ad impression starting February 2027, and a year of Shopping ad data suggests AI Overviews are shrinking the audience that ever sees the ads.

Read the issue →

September 3, 2026 · 7 min read

The answer stopped showing where it came from

Elon University’s survey puts 27% of internet-using US adults in personal conversation with a chatbot and 39% telling it things they would not tell a person, a tracking study of more than two million product listings finds Google’s AI Mode and the ordinary results carousel overlapping on 1.28% of products and naming a different seller first about half the time, Gemini 3.8 Flash reaches AI Mode and the early answers carry no citations, Google confirms a bug unlinking Local Services Ads accounts from Google Business Profiles mid-migration, and Judge Brinkema refuses to make Google sell AdX and orders conduct changes instead.

Read the issue →

September 2, 2026 · 7 min read

Somebody else already filled in your AI settings

A survey of 1,428 people who had been in a legal dispute finds a sixth used a chatbot and some used it to check advice from their own lawyer, Common Crawl reads 584,107 llms.txt files and finds 68.27% written by a plugin and the crawler opt-outs inside them blocking nothing, Anthropic switches on an invisible text watermark worldwide to meet the EU transparency code, Google begins auto-migrating Search campaigns on campaign-level Broad Match and Automatically Created Assets to AI Max through September, and ChatGPT ads hit a $1 billion annualised run rate as self-serve buying opens across 31 European markets.

Read the issue →

September 1, 2026 · 7 min read

The front door started asking who you are

Google is testing a “verifying your request” screen in front of its own results with no sign-in required, a study of 8,000 local searches and 86,645 citations finds AI Mode sending 79.8% of its local citations back to Google Maps while AI Overviews sends 73.5% to business websites, Google’s European search dataset of ranking, query, click and view data now has published eligibility criteria and FRAND pricing, the Search generative AI opt-out control reached every website worldwide on August 31, and Google Ads began labelling a “What customers love” block as AI-generated from store rating reviews.

Read the issue →

August 31, 2026 · 7 min read

Nobody has to leave the answer anymore

Google’s AI Mode starts completing hotel bookings inside Search with Google Pay across ten named partners while the hotel stays merchant of record, a “107 million AI answers” citation study turns out to be a sponsored post with no published methodology, Microsoft requires sign-in for every Copilot user but not for Copilot Search, Google’s site reputation abuse policy stopped applying to EEA searchers on August 30, and Shopping ads moved inline into AI Mode answers.

Read the issue →

August 30, 2026 · 7 min read

Everyone gets footnoted, almost nobody gets named

Brainlabs opens its full 54-client roster showing organic sessions down 10.5% while AI referrals rose 163% and converted at roughly 1.5 times the rate, a Semrush study of 1,458 brands finds legal and healthcare brands earning citations without ever being named, a 129.3-million-citation study puts publishers with an OpenAI licensing deal 48% ahead on ChatGPT, the IRS drops out of Google over its own bot-blocking, and ChatGPT ads go live in India with self-serve access opening September 4.

Read the issue →

August 29, 2026 · 7 min read

The defaults moved and nobody asked you first

Ahrefs reads 145 million search results and finds Reddit in 83.9% of Google’s Discussions and forums box worldwide with three sites carrying 76.75% of its links and no correlation to backlinks, Google confirms AI Overviews now expand into full AI Mode answers by default on some queries, site reputation penalties stop applying to searchers inside the EEA on August 30, new Google tag snippets drop the gtag config command, and Reserve with Google Book buttons reach Search and Performance Max with no opt-in required.

Read the issue →

August 28, 2026 · 7 min read

Everyone sounds certain, nobody can show the work

SE Ranking finds 16.71% of Google top-10 URLs fell past position 100 during the August spam update against a 9.2% July baseline, an arXiv audit of four LLM recommenders finds 11.6% to 38.7% of Yelp businesses named are not in the catalog while stated confidence barely moves, Judge Mehta tells Google the publisher bargain “all seems really unfair,” Google auto-upgrades broad match and automatically created assets to AI Max on September 1, and Demand Gen tests a message thread that opens from a YouTube ad.

Read the issue →

August 27, 2026 · 7 min read

Everything now happens before they reach your site

Google expands Local Services Ads booking from about 20 Reserve with Google partners to more than 500 and switches it on automatically for profiles that already have a booking link, an index built from 4,320 model responses finds 471 brands with high domain ratings and heavy organic traffic that the models barely mention, Google confirms it is routing result clicks through a goto passthrough URL that third-party trackers cannot decode, Google Cloud ships Gemini Enterprise for Legal in preview with four large firms named, and Apple Maps ads go live across the US and Canada with an October 11 deadline on the introductory credit.

Read the issue →

August 26, 2026 · 7 min read

The platforms took over the scorekeeping

Adobe puts AI-referred retail conversion 60% ahead of other traffic after trailing by 128% a year ago, a 5.32-million-result study finds Google trimmed listicles about four points at position one while 92% of those pages also carry Reddit or YouTube, 1.8 million Google Business Profiles favor specific categories over broad ones, Google starts charging for missed Local Services calls on October 1, and ChatGPT Ads begins crediting conversions nobody clicked.

Read the issue →

August 25, 2026 · 7 min read

The dial dropped, the traffic never did

Google confirms a logging error that undercounted Search Console’s Generative AI impressions from August 13, Promptwatch logs Reddit’s ChatGPT citation share falling 86.4% in four days with no explanation, a survey of 602 marketers finds 84% confident and 56% unable to distinguish AI search from SEO, a critical review of 45 GEO studies finds no durable cross-platform effect, and ChatGPT Ads starts reporting view-through conversions.

Read the issue →

August 23, 2026 · 7 min read

Writing got cheap, proof got expensive

Pew scores 490,000 pages and finds signs of AI authorship in over a third of everything published since ChatGPT, a new benchmark puts pages engineered for AI citation at 8.90% of the web and 16.36% of pages modified this year, Google runs Local Services Ads before the badge clears, Googlebot stops repairing malformed JSON-LD, and Microsoft removes Max CPC from new campaigns on October 1.

Read the issue →

August 21, 2026 · 7 min read

Ask again in five minutes, get another firm

A 14,472-citation local study finds Gemini naming the same top business on repeat queries 7.9% of the time against Google’s 90.2%, a preregistered experiment with 1,100 people measures what AI answers do to publisher clicks, Google ships a reader button that feeds AI Overviews, Google Ads auto-migrates search campaigns from September 1, and ChatGPT Ads reaches 31 more countries.

Read the issue →

August 20, 2026 · 7 min read

The audience got warier, the machines got the pen

Semrush and an independent tracker both log the video carousel collapsing on August 12, Pew’s survey of 3,488 adults puts concern ahead of excitement for the first time among under-30s, Cloudflare sets a September 15 default that blocks crawlers training while they index, the IAB publishes AI ad-label standards behind three live statutes, and Microsoft switches AI Max on worldwide.

Read the issue →

August 19, 2026 · 7 min read

Your words arrived, your name did not

A 265-citation passage-level study finds pages that rank first feeding Google’s answer with no credit, a 13.1-billion-event clickstream puts 86% of Google clicks in the discovery column, Search Console turns social accounts into verifiable properties, nine months of tracking put AI Overview misattribution at 22.4%, and Google shortened the Branded Searches window to seven days.

Read the issue →

August 18, 2026 · 7 min read

The controls came off while you were working

Shopify’s data shows AI referrals up 197% while organic search grew 12%, a 40,068-response audit finds a list position alone worth $11 a visit, Google’s August 2026 spam update starts rolling out, language targeting leaves Search campaigns in September, and a New York City Bar opinion sets the default for AI notetakers.

Read the issue →

August 17, 2026 · 7 min read

The shortlist was written before you walked in

ChatGPT writes brand names into its own search queries before it fetches a page, Google’s Gemini app passes a billion monthly users with 63% talking instead of typing, crawler data shows ChatGPT’s fetch bot reading disallowed pages, Microsoft Clarity puts a number on scraping versus referrals, and Google Ads starts narrating its own results.

Read the issue →

August 16, 2026 · 7 min read

The chat window put in a front desk

Anthropic starts watermarking Claude’s text, Google expands Limited Ad Serving with impression caps for unverified advertisers, YouGov puts AI names in its brand rankings while trust sits at 28%, a joke cat file exposes the evidence bar behind AI-visibility tactics, and ChatGPT’s ad platform flips a data default tomorrow.

Read the issue →

August 15, 2026 · 7 min read

Your next client arrives already decided

Shopify’s Q2 data shows AI-referred shoppers converting at twice the organic rate, a Scrunch panel study finds AI Overviews mostly land where clicks were already scarce, Google is putting AI-generated images inside AI Overviews, and two Google Ads controls go automatic in September: language targeting and legacy campaign structures moving to AI Max.

Read the issue →

August 14, 2026 · 7 min read

Your citation works even when nobody clicks

A 900-adult browsing panel finds only about 1% of AI Overview visits click a cited source, Demandbase counts ChatGPT referrals to business sites up fourfold in a year, OpenAI is tucking ChatGPT’s sources behind an extra click, Google Ads added reporting-only new-customer tracking, and Apple is negotiating pay-per-use content deals for the new Siri.

Read the issue →

August 13, 2026 · 7 min read

A stranger’s listicle is speaking for your firm

Google’s local AI Overviews cited one attorney’s self-promoting roundup for a whole list of law firms, a 35,000-citation analysis put community sites four times ahead of publishers, nearly 300 French newspapers filed a competition complaint over AI Overviews, Search Console’s generative AI report reached far more properties, and Performance Max added a local customer toggle.

Read the issue →

August 12, 2026 · 7 min read

The settings you ignored started moving themselves

Google confirmed AI Mode buttons on its home page, a 45-study survey finds no proven durable AI-visibility technique, Meta’s documented crawler points at a search index of its own, two Google Ads settings auto-upgrade to AI Max in September, and ChatGPT ads went live in five more countries.

Read the issue →

August 11, 2026 · 7 min read

Ad money buys the slot, not the answer

SE Ranking found ChatGPT ads on 25.94% of commercial prompts but the advertiser cited in the answer above only 3.63% of the time, Google’s generative AI report went live for every Search Console property, Yelp put restaurant booking inside ChatGPT, and OpenAI switches on automatic pixel matching August 17.

Read the issue →

August 10, 2026 · 7 min read

Paperwork quietly became the new ranking factor

A census of 4,776 venues found 85.6% are never recommended by AI assistants, People Inc’s sessions fell 22% while session revenue fell 1%, OpenAI changed the default ChatGPT model for free users, and Google will demand proof of government authorization from October 5.

Read the issue →

August 9, 2026 · 7 min read

The shortlist quietly replaced the results page

Hello! rebuilt after losing Google Discover, 50,000 open fan-out prompts show two question types carry 97% of brand mentions, Google Maps starts completing orders, and OpenAI switches on automatic form matching on August 17.

Read the issue →

August 7, 2026 · 7 min read

A Machine Now Picks Which Page Speaks for You

Shopify says AI shoppers arrive closer to the decision, server logs show ChatGPT reaching for English pages 2.6x more often than Google, and two hard dates land — Google Ads bidding on August 17, the end of Google Assistant on September 4.

Read the issue →

August 6, 2026 · 7 min read

The Ad on Top, the Verdict Below

A sponsored ad and Google’s own summary named different companies on the same screen, a six-month study of 1,094 categories separates being cited from being recommended, and Google’s government-services ad policy tightens on October 5. Five things that moved in AI search.

Read the issue →

August 5, 2026 · 7 min read

The referral click became a rounding error

ChatGPT sends an outbound click in 5.2% of sessions, a controlled experiment nearly doubled machine task success on identical content, and two Google Ads settings flip to automatic in September.

Read the issue →

August 3, 2026 · 7 min read

The Week the Answer Box Grew a Cash Register

Consumers stopped fact-checking the robot, ChatGPT bought a library card for local reviews, and the answer box got a bid button — five developments, all sourced.

Read the issue →