Good Monday Morning

It’s June 2nd. The Hajj begins Wednesday with an estimated two million Muslims expected to travel to Saudi Arabia. Extreme heat last year killed more than 1,300 people.

Today’s Spotlight is 1,058 words, just over 4 minutes to read.

3 Headlines to Know

Lopez Sued Over Her Own Paparazzi Pics

Jennifer Lopez faces a new lawsuit for posting photos of herself in designer outfits to social media without the photographers’ permission. Remember, photographers and not subjects, own the photos.

Grammarly Nabs $1B Without Giving Up Equity

Grammarly just scored $1 billion in nondilutive funding, showing that AI giants like ChatGPT and Claude have not pushed it off the map as the company doubles down on marketing and acquisitions without losing any equity.

Gemini Can Now Watch Your Google Drive Videos

Google’s Gemini will now scan your Drive videos and spit out instant summaries or answers, so you can skip the playback and get straight to what matters.

Most Americans See AI Mentions, Few Dig Deeper

By The Numbers

George’s Data Take

Despite 93% seeing AI mentioned online, just 8% actually read a news article that talked about AI in depth. Everyone hears the buzz, like a brood of cicadas, but fewer than 1 in 10 dig deeper to find out where the sound is coming from.

Good & bad: You’re early still if you’re integrating AI, but you’re going to have to educate as you go.

This gets critical in the next piece.

Duolingo’s AI Move Triggers Viral Backlash

Running Your Business

When Duolingo announced it would lean on AI and phase out contractors, users revolted with one-star reviews and TikTok funerals, proving you can’t meme your way out when fans feel wronged.

Silver Beacon Behind the Scene

Even if you’re a tech company, you cannot let the nerds drive the message. The world doesn’t want to know about people losing work to LLMs or generative AI.

That’s a message for your investors, not your customers, and I know you know not to make those the same message. 

Your Tattletale Car

Image by ChatGPT, prompted by George Bounacos

The conveniences inside our cars have also enabled sophisticated data tracking to accompany the constant surveillance of our roads. 

WIRED recently reviewed police training documents teaching agencies how to access data from connected cars. Their biggest data jackpot is from subscription services who often supplement low monthly fees with a booming data business even when vehicles aren’t mapped by our phones, tollbooths, and traffic cameras.

Your Car’s Close-Up

Don’t ignore license plate cameras, though. Last month a county sheriff in Texas performed a nationwide search of more than 83,000 automated license plate readers (ALPR) throughout the country. The criminal he was searching for was a woman he accused of performing a self-administered abortion. 

The sheriff told 404 Media that he searched nationwide to “[hit] everything, every possibility.”  But while most forms of abortion are illegal in Texas, the sheriff was able to search for the woman’s car throughout the country, including states where abortion is a fundamental right. 

The ALPR networks continually scan passing vehicles for their license plates, make, model, and color. Coupled with date and time stamps, they’re a nationally accessible permanent record of where the vehicle has been. 

Oregon Keeps Trying To Shut It Down

Car privacy data is big business too. Auto insurers can set rates based on where cars are seen, how often they’re driven, and how fast. Life insurance rates, credit card interest, and even whether you qualify for a mortgage are all informed by behavioral models that are influenced by vehicle data.

And you’re helpless to change that because car companies collect and sell your driving data constantly, but they won’t tell you who they’re sharing it with. When Oregon passed a law allowing residents to request a list of all companies that get their personal data, nearly 400 people asked Privacy4Cars to file these requests with car manufacturers. Despite the law requiring it, not a single car company provided the list.

Oregon’s legislature amended the law last week to strengthen it, but there are still plenty of gray areas and loopholes. The need for comprehensive federal standards is urgent, but states looking out for their citizens could at least follow Oregon’s example in the interim.

Sun-Times Prints Fake AI Summer Reads, Goes Viral

Practical AI

The Chicago Sun-Times ran an AI-generated Summer reading list packed with books and quotes that do not exist, sparking embarrassment, newsroom outrage, and a crash course in the dangers of not fact-checking.

New DMV Text Scam Hits Phones Nationwide

Protip

Fake DMV texts are now sharper and harder to spot thanks to AI, threatening license suspension to scare drivers into handing over personal info as scammers get smarter. I’m getting several of these each week.

Trump’s Harvard Math Claims Flunk the Facts

Debunking Junk

Trump claimed Harvard teaches remedial arithmetic like two plus two, but Harvard’s entry level math course is calculus and its new support class helps students with calculus and higher math. 

I checked the data. Of 800 possible points on the math SAT, the national average score is 521. The average at Harvard is 790. The top quartile has perfect math scores and even the lowest quartile averages 760.

Mads Mikkelsen and Campari’s Cannes Spot

Screening Room

Caltech’s Smart Bandage Spots Infections Before Symptoms

Science Fiction World

A new smart bandage can sense signs of infection and inflammation days before symptoms appear by using real-time fluid analysis and AI to predict healing and offering patients a testing lab on their skin. There is no word on what the bandage AI thinks about this summer’s reading.

New Enzyme Doubles Efficiency of Turning Plant Waste into Biofuel

Tech For Good

Scientists discovered a natural enzyme named CelOCE that unlocks cellulose in agricultural waste much faster than before, which could mean cleaner and more efficient biofuel production.

Roadtrip With Hundreds of Strangers

Coffee Break

Neal Argawal, my favorite web author, is back with Internet Trip,  a crowd based app where you join with hundreds of others to steer a car through Google Street View. They’re deep in Nova Scotia as I write this, lazily making their way towards Halifax and sometimes honking the horn for no reason. 

Sign of the Times

Good Monday Morning!

It’s May 19th. Housekeeping: we’re off next week for Memorial Day and back in your email on June 1 at 6 a.m.

Today’s Spotlight is 895 words, just over 3 minutes to read.

3 Headlines to Know

Verizon Drops DEI to Secure $20B Fiber Deal

Verizon Won FCC approval for its $20 billion merger with Frontier only after agreeing to drop DEI initiatives and teams from its organization.

Google Bets Big on Next-Gen Nuclear Power

Google backs three new nuclear sites to fuel its AI data centers with low-carbon power.

Netflix Will Use AI to Disguise Ads as Content

Netflix will let advertisers blend their spots into shows and movies to make commercials on non-premium tiers look less like ads.

Younger Consumers Don’t Want DEI Cuts

By the Numbers

George’s Data Take

Promoting diversity, equity, and inclusion was once the safe move, but can now put companies in the crossfire between federal demands and customer desires. 

Notice the gap: 29 percent of boomers say canceling DEI conflicts with their values, but only 9 percent mention DEI directly. For many, “DEI” signals deeper political divides in a fiercely tribal time.

VPN Firm Kills “Lifetime” Subscriptions Without Warning

VPNSecure’s new owners say they did not know about the lifetime accounts before buying the company and shut them down, leaving longtime customers cut off.

Silver Beacon Behind The Scenes

The new owners first claimed they did not know about the lifetime accounts, then argued they only bought the assets, not the company. Either way, this was a spectacularly bad move.

At best, a screwup like this means costly and time-consuming complaints. At worst, it could bring lawsuits or trigger regulatory action.

Image by ChatGPT, prompted by George Bounacos

Big Change

Artificial intelligence is being chased by every major search player.

The promise: synthesizing and summarizing information from across the web, not just linking to it. Google and OpenAI’s ChatGPT are both rolling out major upgrades that push the web closer to answer engines instead of just lists of links.

Google Gets Chatty

AI Mode is not just another AI Overview. For months, Google has shown AI answers at the top of search results, but now AI Mode brings a persistent, interactive answer bar to mobile. 

Instead of only showing summaries after you search, you get a chat-style box that you can open, ask follow-ups, and see responses as you type even before you hit enter. This makes mobile search feel more like messaging and less like old-school searching. The feature is rolling out through Labs to some Android and iOS users, and it is noticeably more conversational than the previous AI tools.

ChatGPT Builds on Bing

OpenAI’s ChatGPT has taken a big step by expanding local and product search. It starts with data from Bing but layers on its own AI analysis, so it is not a copy-paste job. 

Microsoft’s thirteen billion dollar investment in OpenAI makes this deep connection possible. If you run a business, you cannot afford to ignore your Bing business profile, because ChatGPT uses it as a foundation for local results.

Local Search is Huge

Stat to know: Nearly half of all Google searches are for local information. That means the AI tools that surface business info, 
reviews, and directions are now the front door for customers.

Product Search Goes Personal

ChatGPT’s new shopping features deliver product recommendations, images, reviews, and links with no ads (yet) cluttering the results. Rankings are based on relevance and user needs, not pay-to-play.

The Takeaway

AI search is less about keywords and more about quality answers and trust. Businesses must keep their info up to date everywhere, especially on Bing and Google, to show up in these new AI-driven results.

Students Add Typos to AI Essays to Fool Detectors

Practical AI

College students are dodging AI plagiarism checks by making chatbot papers look more human using deliberate mistakes and clumsy prompts.

OpenAI Lets Users Export Research as Polished PDFS

Protip

ChatGPT users can now save and share reports in PDF format, making it easier for businesses to distribute and verify AI research. This is especially helpful for the company’s new multipage output when deep research mode is used.

Trump Posts Video Accusing Clintons of Murder

Debunking Junk

President Trump published a debunked video online Saturday that falsely claims his predecessor and his former chief rival are tied to multiple deaths of their employees or potential rivals.

Michael J Fox and Proud Canadian Create Proud Canadian Spot

Screening Room

Facemask Sensors Detect Kidney Disease on Your Breath

Science Fiction World

Researchers built disposable masks with embedded sensors that can identify chronic kidney disease with 93% accuracy by analyzing what you exhale.

Custom Gene Editing Saves Infant With Rare Disease

Tech For Good

with 93% accuracy by analyzing what you exhale CRISPR therapy tailored for one baby’s mutation, opening new hope for person-specific treatment of rare genetic disorders.

AI Tool Guesses Your Face’s Real Age

Coffee Break

FaceAge uses millions of images and skin markers to estimate how old you look and gives you advice for healthier skin. Yes, you have to give them or take a picture. 

Sign of the Times

Good Monday Morning

It’s May 5th. Wednesday is when REAL ID driver’s licenses or a passport are required to board an airplane. This DOT site shows you what is needed in each state to get yours.

Today’s Spotlight is 838 words, about 3 minutes to read.

3 Headlines to Know

Google Play Loses Nearly Half of It’s Apps

Google removed 1.8 million apps in the last year after tightening rules to push out low quality and policy violating content.

Law Firm Asks Supreme Court to Ignore Search Ad Lawsuit

A personal injury firm told the Court a competitor’s trademark fight over search keywords is nothing new and already settled by lower rulings. At issue: whether you can use a competitor’s name to trigger ads form your website.

Wikipedia Starts Using AI to Help Editors

The encyclopedia site will use generative AI for grunt work like translation and research so human editors can focus on quality and moderation.

Streaming Hits 43% of TV Views

By The Numbers

George’s Data Take

These big numbers show how much streaming continues eclipsing traditional and broadcast, now accounting for about the same amount of viewers. Tubi and Roku may not have the cachet of NBC or Warner Brothers, but that’s where eyeballs and advertising targets are. 

Shortcuts Can Be Expensive

Running Your Business

Attorney General Pam Bondi told TV viewers and then a Cabinet meeting that fentanyl seizures during the first 100 days of Trump’s presidency “saved 258 million lives.” 

Luckily, the fentanyl crisis isn’t bad enough to have threatened 75% of the U.S. population in just 3 months. Here’s what happened.

Silver Beacon Behind the Scenes

Analysts do more than make spreadsheets with pretty charts. They delve into how the numbers are created and what they represent. They’re the detectives behind the person presenting the data, but that person still has to understand what the numbers represent. 

In this case, DOJ summed the amount of fentanyl seized, divided by a lethal dose, and came up with a huge number. The problem is that 258 million Americans don’t use fentanyl.

The arithmetic was correct. The analysis was horribly wrong. 

AI is Flooding Bookshelves With Trash. There’s No Solution in Sight

Image by ChatGPT, prompted by George Bounacos

Thousands of AI-generated ebooks are swamping public libraries, online bookstores, and search results. These aren’t clever novelties. They’re confusing, plagiarized, and sometimes dangerous.

404 Media’s latest investigation found entire shelves of public library offerings via the app Hoopla are full of bot-generated books that are often designed to trick people into thinking they were authored by humans. It’s the book world’s version of Spotify promoting AI-made songs from fake artists, hoping you won’t notice.

One reader spotted over a thousand books on tatting lace that were algorithmic nonsense. Then came a bigger problem: AI books giving medical advice.

The Guardian found dozens of self-published Amazon titles offering ADHD guidance written entirely by AI. The books featured a grab bag of fraudulent writing that careened from invented experts to fake quotes to dubious and harmful claims. 

All were written with the apparent confidence of well-trained professionals using appropriate jargon and trustworthy phrases because that is exactly the type of task that large language generative AI models excel at.

Silver Beacon’s Take

Sounding authoritative has never been easier. 

Self-publishing once carried a stigma. Now, plenty of credible authors go that route. But the barrier to entry has dropped to near zero. That opens the door for grifters, copy-pasters, and anyone else with a text generator and a free Canva account.

The FTC is pressuring Amazon to address the surge of AI-generated books being sold on its platform, especially those with misleading health information. Critics blame Amazon’s self-publishing model for incentivizing lower quality books to be churned out at scale while not including vetting or other editorial safeguards. 

Platforms like Amazon and Hoopla still haven’t caught up. Some titles get removed if flagged, but there’s no systemic fix or oversight in place. Amazon downplays the problem. Hoopla says it’s reviewing things.

In the meantime, niche hobbyists, parents, and readers looking for answers are wading through a swamp of garbage. And the AI engines keep churning even for authors who may know nothing about the subject.

OpenAI Pulls ChatGPT Update After Users Mock Praise Overlord

Practical AI

Open AI CEO Sam Altman says a recent update made the chatbot annoyingly flattering, so the company is rolling it back to be less of a cloying yes-bot.

Spotify Adds Option to Turn Off Smart Shuffle

Protip

A new setting lets users disable Spotify’s Smart Shuffle, the feature that inserts suggested songs into personal playlists without asking.

Trump Keeps Inventing Cheap Gas Prices

Debunking Junk

The president now claims gas is $1.88 in three states, but the real average is $3.19. Gas Buddy and AAA say that no state or station comes close to his numbers.

Olivia Colmon Inspects Crumpets

Screening Room

Driverless Trucks Begin Long Haul Routes in Texas

Science Fiction World

Aurora’s self-driving rigs are now delivering freight between Dallas and Houston using sensors and cameras instead of a human behind the wheel.

New AI Tool Finds Tumor Flaws For Personalized Treatment

Tech For Good

Cambridge researchers built an algorithm that scans tumor DNA for hidden repair defects so doctors can match patients with therapies like immunotherapy that target their specific cancer.

This Music DNA Project Visually Maps How Songs Relate

Coffee Break

The Pudding charts thousands of tracks by energy and complexity—placing “Bohemian Rhapsody” near classical opera and “WAP” next to punk.

Sign of the Times