Political Deception Online – Spotlight #464
News

Good Monday Morning


It’s June 12. The Fed wraps up its meetings on Wednesday and is widely expected to defer raising rates after ten straight increases.

Today’s Spotlight is 1,210 words–about 4 1/2 minutes to read.

3 Stories to Know

1. Federal deposit insurance does not cover money stored in non-bank apps like PayPal or Venmo, warns the Consumer Financial Protection Bureau. In the U.S., 85% of adults aged 18-29 use peer-to-peer payment apps.

2. A toner manufacturers’ trade associationhas complained to the Global Electronics Council about HP’s new policy that prohibits consumers from using third-party toner. Consumers who accept HP’s offer of 6 months free ink for a new printer also opt into a program that requires using only HP ink for the life of the printer.

3. Big Tech companies still block their software from labeling any image as a gorilla eight years after Google’s app mislabeled two Black people as gorillas. As part of its test, the New York Times used images of people, animals, and objects to test the software of Google, Apple, Amazon, and Microsoft. Gorilla identification was not done by any of the four companies, although Google and Apple were able to identify the other animals.

Spotlight on Political Deception Online

With generative AI, creating political deception online has become more accessible to millions of people who don’t need advanced equipment or training to do so. 

Within hours of Gov. Ron DeSantis’ announcement to run for president, a PAC allied with him faced criticism for an ad featuring fake fighter jets flying over his speech. The DeSantis campaign also posted a video last week with doctored images depicting Donald Trump hugging Dr. Anthony Fauci three times.

Researchers have anticipated escalating political deception online for years. Even though doctored videos were used during the 2016 and 2020 presidential elections, the upcoming election represents the first time the electorate faces a threat when virtually anyone can create a look-alike video.

Reversing the Past

In previous elections, Facebook, YouTube, and other Big Tech firms attempted to police misinformation and disinformation. Both companies changed internal algorithms to reduce the amount of political content shown to users. For Facebook especially, its focus on private groups proved detrimental to that effort. Last year, Mark Zuckerberg ended the program after political content during the 2022 election was shown only half as much as during the 2020 election.

However, Facebook’s advertising algorithms still allowed researchers to successfully submit ads calling for violence against election workers. Facebook’s system allowed 15 of twenty ads to proceed despite YouTube and TikTok halting them.

YouTube has reversed itself as well. The company said that videos with false complaints about elections will remain visible on the site. Content that misrepresents voting logistics and eligibility will continue to be removed from YouTube. 

Violent Language Online Increasing

According to authorities, extreme expressions of political opinion online have increased since last week’s indictment of Donald Trump by a federal grand jury. GOP members of Congress are contributing to the extremist talk, including Rep. Clay Higgins (R-LA) who tweeted military-style information to followers and Rep. Andy Biggs (R-AZ) who tweeted, “We have now reached a war phase. An eye for an eye.”

A doctored video sent by Donald Trump himself two days ago ratcheted up tensions yet again. The former president posted a video of himself hitting a golf ball mashed up with sound effects and a video of President Biden falling while riding a bicycle as if he had been struck by the ball. The video was reminiscent of Trump posting a picture of himself with an extended baseball bat next to an image of New York District Attorney Alvin Bragg, who had recently announced state charges against Trump.

Practical AI

Quotable: A UK firm found only half of adults could tell the difference between advertising emails written by generative AI or a human copywriter.

Noteworthy: An attorney who claimed to use ChatGPT for legal research is the latest person to run afoul of model glitches called “hallucinations.” In his research, the software invented previous cases, which the attorney then included in a federal court filing. The judge in that case has ordered a hearing to review possible sanctions.

In a completely unrelated case, a contributor to a publication covering guns asked ChatGPT for information about a radio host. ChatGPT erroneously posted that the radio host had been accused of misappropriating funds. The radio host is now suing ChatGPT developer OpenAI for libel.

Tool of the Week: Adobe’s Generative AI Fill extends images pretty easily. By now you’ve probably seen images of the Mona Lisa or other famous pictures with surrounding scenes. The Verge has some interactive samples so you can see for yourself what the fuss is about.

 Waiting in the Wings

  • How algorithms are automatically denying medical claims
  • Amazon’s data about you expands beyond shopping

Put your email address in the form at this link and you’ll get a free copy of Spotlight each Monday morning to start your week in the know.

If you’re already a free subscriber, would you please forward this to a friend who could use a little Spotlight in their Monday mornings? It would really help us out.

Trends, Spends & TikTok

 Did That Really Happen? — Climate Change Conspiracy Theory Isn’t True

Podcaster Joe Rogan is being criticized again after a segment on his show included a nonsensical climate change theory that claims the Earth’s magnetic poles reverse themselves for 6 days every 6,500 years and then conveniently revert on the seventh day. That segment has now found its way into various viral spots, including TikTok.

Following Up —  Airbnb Sues New York

We wrote extensively about Airbnb three months ago and suggested that more cities would be cracking down on the short-term rental platform. While that appears to be happening, Airbnb filed suit against New York City earlier this month for what it called “extreme” and “oppressive” rules that run counter to federal law.

Protip — Changing Android’s Keyboard Size

If you’re missing the correct keys on your Android phone’s screen more often than you would like, this ZD Net walkthrough on changing its size is just what you need.

Screening Room —  Apple’s iPhone Privacy

Apple promotes the privacy settings of the iPhone’s Health app in this fun new spot.

Science Fiction World —  Brain Implants 

This technology is coming faster than  perhaps anyone dared dream, but an experimental brain implant that boosts nervous system signals is working. The BBC has coverage of a 40-year-old man who is paralyzed, but can move his legs via the implant. It sounds farfetched, but the Lausanne University research has been published in Nature.

Coffee Break — Ride the Space Elevator

Climb in at ground level and ride this great Neal Agrawal interactive past birds, kites, and all the way up to space.

Articles

Good Monday Morning


It’s May 22. We’re off next week for Memorial Day. See you on June 5.

Today’s Spotlight is 1,173 words–about 4 1/2 minutes to read.

image by MidJourney 5, prompted by George Bounacos

Spotlight On … Federal Surveillance Programs

A steady drumbeat of anti-privacy federal surveillance headlines emerged last week about programs at FBI, HUD, and the Post Office. The programs are similar to one we learned about in January regarding surprising federal surveillance initiatives.

For Everyday Money Transfers

In January, we learned that more than 600 law enforcement agencies can access a database that details money transfers between Americans and people in more than 20 countries. Those transfers include the full name and address of the sender and recipient for money orders and other electronic transfers well below the $10,000 threshold, often triggered by as little as $500.

Important quote: “Courts have held that customers using money transmitter businesses do not have the same expectation of privacy as traditional banking customers,” an Arizona Attorney General statement to the WSJ.

In Public Housing Communities

Public housing funds earmarked to combat crime are being used to monitor resident behavior and recommend evictions for minor violations. Transgressions such as taking a cart from a laundry room or violating overnight guest rules are monitored by cameras that include facial recognition and advanced software.

Important quote: “People choose to get evicted by their actions.” Melody McClurg, executive director of the Jefferson Metropolitan Housing Authority in Steubenville to the Washington Post

At the Post Office

A bipartisan group of U.S. Senators last week accused the U.S. Postal Service of “unchecked government monitoring” of Americans’ mail. Post offices collect information upon written request by a law enforcement officer–no court order is necessary. All the information on the envelope’s outside is included. In their letter, the senators claim tens of thousands of Americans and businesses are monitored.

Important quote: “A mail cover is an investigative tool used to record data appearing on the outside of a mailpiece. Law enforcement agencies use this information to protect national security; locate fugitives; obtain evidence; or help identify property, proceeds, or assets forfeitable under criminal law. A mail cover is justified when it will further an investigation or provide evidence of a crime.” — USPS description

In Secret Databases

A recently unsealed court document shows that the FBI used a federal surveillance database to monitor individuals who protested for racial equality, illegally entered the Capitol during the January 6 insurrection, or donated to a congressional candidate. The database is under the control of the Foreign Intelligence Surveillance Court, which threatened last year to restrict FBI access to the data.

Important finding: Court documents show “an FBI analyst conducted a batch query for over 19,000 donors to a congressional campaign” because the analyst said the campaign was a target of foreign influence, reports the Washington Post.

3 More Stories to Know

1) A First Amendment battle is shaping up in Montana where Gov. Greg Gianforte signed a bill to ban TikTok in the state. The bill takes effect in January and prohibits app stores from making the software available and the company from operating in the state.

2) 30-Second unskippable ads can now be purchased on YouTube. Calling the format “a staple of the TV industry,” YouTube said that the ads would be used only when content is shown on a television and not on phones or computers.

3) Angi company HomeAdvisor has agreed to pay $7.2 million to home services companies it defrauded by misrepresenting the consumer leads it sold them.

Did That Really Happen? — Biden Graduation Video Manipulated

A video of President Biden attending his granddaughter’s graduation has been digitally manipulated to include a crowd chanting “Fuck Joe Biden” in time to a marching band. The hoax video has been viewed more than one million times.

Following Up —  TSA Rolling Out New Program

Travelers at 16 airports will have the opportunity to take a selfie at a kiosk and then have their id scanned without having to rely on an officer to match their id and face. An agent is available, and travelers can opt out.

Protip —  Revisit WhatsApp Privacy

Wired updated their privacy walk-through of WhatsApp so if that’s one of your communications channels, have a look through this explainer.

Screening Room —  Augmental MouthPad

This unobtrusive mouthguard-like device is a game-changer for adapting tech devices for people who have disabilities.

Science Fiction World —  Your DNA, Pulled From Thin Air

Environmental researchers have long filtered human DNA from their work. Wildlife researchers and botanists use environmental DNA after scrubbing out human contamination. Now a bioethical issue is being raised: what should be the rules for how residual human DNA is used when it’s available to anyone with the right equipment?

Coffee Break — Same Place Names As Viewed Throughout The U.S.

This interactive database allows you to enter a place name and learn which town is being referred to throughout the U.S. 

For example: most of the country is referring to the Fairfax in Virginia, but people in Kentucky, Ohio, and Indiana likely mean Fairfax, Ohio, and people in SC and eastern Georgia probably mean Fairfax, SC.

It’s great fun to play with.

Sign of The Times

News

Good Monday Morning


It’s May 15. Open AI co-founder and CEO Sam Altman testifies before a Senate Judiciary subcommittee tomorrow, the first time the executive whose firm launched ChatGPT has appeared before Congress. Don’t expect legislators to show too much knowledge. The House only purchased 40 ChatGPT licenses and those were bought three weeks ago.

Today’s Spotlight is 918 words–about 3 1/2 minutes to read.

Image by MidJourney 5.1, prompted by George Bounacos

Spotlight On … Increasing Ransomware Attacks

Ransomware attacks are increasing in 2023, thanks in part to a vulnerability in widely used commercial software that was exploited against more than 100 companies in 30 days. That’s not the only reason; ransomware attacks also increased in January and February compared to last year.

The First Wave of Ransomware Attacks

The Toronto city government confirmed it had been attacked three days after reporters from TechCrunch contacted the administration. Other hacking victims include:

  • 1 million Community Health Systems patient records
  • 140,000 Social Security records from Hatch Bank
  • Hitachi Energy employee data
  • Internal data from cybersecurity company Rubrik

These aren’t the only recent ransomware attacks. In March, personal records about student mental health, suspension reports, and even notes about rape and child abuse of Minneapolis students were published online after the city refused to pay a $1 million ransom.

Hackers attacking Virginia’s Bluefield University took their case directly to the university community last week. After breaching the school’s computers, the hackers used the college’s alert system to warn students and faculty that their records would be published if the university didn’t pay.

Employees Are An Attack Vector

There will likely be more attacks as hackers target tech employees.

More than 27 million passwords purportedly belonging to Fortune 1000 company employees are available on the dark web. We also learned this week that a third-party support agent working for Discord was hacked.  That attack led the network to inform users that hackers may have downloaded their email addresses and attachments.

Meanwhile SchoolDude software acknowledged last Friday that hackers have stolen nearly three million user records of school administrators, principals, and campus maintenance workers who submit maintenance and repair requests.

2) The New York Times and Google reached a $100 million, three year deal that will feature Times content on Google as well as subscriptions and ad products. Other publishers with similar Google deals include The Guardian and The Wall Street Journal.

3) Separately, Google announced last week during its I/O Event that it will bundle its generative AI software called Bard into Search and Google Docs. The company is also accelerating its Immersive Maps view and announced new tablets, a budget phone, and a luxury folding phone that will cost $1,800.

Practical AI

Photos published on social media by Amnesty International to highlight police crackdowns on protests in Colombia were AI generated fakes. Amnesty said that it used images that it labeled as AI-generated to protect protesters from retribution.

The UK government is adding £100 million in funding to establish a public-private AI task force to boost the country’s competitiveness and ensure “safe and reliable” use of the technology. This is the government’s second major financial commitment and follows a £900 million investment for computing hardware and AI research.

Tool of the Week: Google has made samples of its MusicLM program publicly available. The software creates new music from text prompts.

 Waiting in the Wings

  • How algorithms are automatically denying medical claims
  • Amazon’s data about you expands beyond shopping
  • Big Tech has a 2024 election issue it hasn’t quite solved 

Put your email address in the form at this link and you’ll get a free copy of Spotlight each Monday morning to start your week in the know.

If you’re already a free subscriber, would you please forward this to a friend who could use a little Spotlight in their Monday mornings? It would really help us out.

Did That Really Happen? — Crime Not Lowest on Mother’s Day

The meme’s timing was perfect. Unfortunately, no one has proven that it’s true. Snopes explains why.

Following Up —  TurboTax Refunds On The Way

We’ve been writing for years about how Intuit’s TurboTax steered consumers away from its participation in the IRS’ Free File program.  Some of those government-filed suits are now putting a little cash in the pockets of more than four million filers who the company misled. Those refunds are averaging about $30.

Protip —  Take a Screenshot of Streaming Video

The Verge has published a walkthrough showing you how to change your browser settings to take a screenshot of Netflix, Amazon Prime, or other video streaming services playing on your device.

Screening Room — Dove Real Virtual Beauty

Science Fiction World —  Lawnmower Makes Room for Bees, Butterflies

We’ve been learning a lot about how lawns can harm native bee and butterfly populations. Now the Roomba-like Husqvarna lawnmower will cut the grass while maintaining 10% of the area for those pollinators. You can see the lawnmower in action here.

Coffee Break — This Bear Knows What He Wants

You definitely want to see this Colorado Parks & Wildlife video showing a bear properly opening a truck door, dumping the stuff that he didn’t want, and loping off with a picnic basket a bag.

Sign of the Times