AI Cheating You Should Understand – Spotlight #452

Good Monday Morning

It’s February 27th. Stop using your air fryer if you own one of the two million that have been sold by Cosori. Their website has details.

Today’s Spotlight is 960 words — about 3 1/2 minutes to read.

Spotlight On …  AI Cheating

As technology advances, AI cheating – making false documents, taking tests, or manipulating video footage – is a growing concern. The potential for AI to be misused grows as it gets more sophisticated, and society is ill-prepared.

Not Photographs, But Still Art

Most focus is on tools such as ChatGPT and Bing’s new chatbot search engine, but this month Jos Avery’s photography secret emerged. He maintained an Instagram following based on his unique portrait photography. But instead of a camera, he used Midjourney, a publicly accessible image model and edited the results. Avery claims he typically creates 85 images per editable image he then manipulates in software such as Adobe Lightroom or Photoshop. Many photographers are furious, but as one observer pointed out, he is simply another artist.

Yet Avery was perpetuating AI cheating, potentially taking someone else’s work and using it to create his own art. Many photographers consider that copyright infringement. Stock image giant Getty Images has accused Stability A.I. of using more than 12 million Getty photographs to train its artificial intelligence image generator, Stable Diffusion.

Video Challenges Include Porn 

AI generative models can create realistic-looking video of someone doing or saying something they never did, or manipulate an existing video to change its meaning. That can be a boon for a movie or television project with the actor’s approval. Harrison Ford, 80, told Variety that his upcoming fifth Indiana Jones movie includes footage of him manipulated to make it appear as though he hadn’t aged since he last played the role decades ago.

Deepfakes and AI-generated videos continue to challenge fact checkers and historians. With deepfakes, videos appear real but are actually manipulated to show something that didn’t happen. The technology has been used to create misleading videos of politicians, celebrities, and ordinary people.

Recent years have seen a rise in the popularity of deepfake porn. Through technology, porn producers can create videos apparently featuring celebrities and other public figures, but increasingly also abusing women who aren’t in the spotlight.

This technology, according to activist Mia Landsem, has made it easier for content creators to integrate the faces of non-consenting women into pornography. Those videos further objectify and dehumanize women, often causing psychological distress and reputational damage. One rare government agency that helps, The UK Revenge Porn Helpline, logs thousands of calls each year.

Tricking a Bank’s Voice Authentication

Even audio fakes are causing AI cheating problems. Voice authentication systems are not as secure as other biometrics, leaving criminals an opening using synthesized voices. Writing in Vice last week, Joseph Cox demonstrated how he tricked his bank, which uses voice authentication, to gain access to his account using a synthesized version of his voice. Cox says he needed only five minutes of audio and some software.

3 More Stories to Know

1)  TikTok parent company ByteDance has launched Lemon8 in the US and UK.  The app is a mix of Pinterest and early-Instagram. Videos aren’t allowed, and there’s still plenty of empty categories. The app was downloaded one million times when it was released in Japan last summer.

2) The Justice Department accused Google of destroying records in the government’s antitrust suit against the company. Google employees routinely used an instant-messaging app that deleted chats after 24 hours while discussing sensitive company information. 

3) 14% of Gen Z (aged tweens to mid-20s) are getting their news information from TikTok according to a new analysis by Morning Consult. Nearly half of Baby Boomers and Gen X said they get most news information online from YouTube.

Trends & Spends

Did That Really Happen? — Ted Cruz & the Xbox

Sen. Ted Cruz (R-TX) stoked more culture wars by saying that “they” were gunning for gas stoves, coffee, and your Xbox. The short version is that Microsoft when the game console updates to be more eco-friendly. Poynter does a fabulous job demonstrating how they fact-checked this misleading statement. I can’t recommend it enough to you.

 Following Up — Prompt Engineers

We’ve written so much about generative AI for a couple of years now that it’s sometimes hard to remember that half of adults don’t know (or maybe care) what it is. Meanwhile, the hot new trend in tech job circles might be “prompt engineers”–people who have the skills to devise unique, effective prompts that coax generative AI models to output something close to what was hoped for or planned.

Protip — Don’t Accidentally Share Your Location

Not sharing your location with every app maker and their marketing partner can be inexplicably difficult. Wired gives some great advice about how to protect yourself.

Screening Room —  Canada’s Humanitarian Coalition

Science Fiction World — Starlink’s Satellite Broadband

It’s touted as broadband anywhere on land for about $200 per month. That’s the price of SpaceX’s Starlink service. The company also has maritime products starting at $5,000 monthly and aviation products at $25,000 monthly.

Coffee Break —  What Historical Icons Might Look Like Today

This artist disclosed that the images are doctored with software. See Al Capone, Ben Franklin, and Babe Ruth among other historical icons reimagined in today’s time. 

Sign of the Times

Leave a Reply

Your email address will not be published. Required fields are marked *
You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>

This site uses Akismet to reduce spam. Learn how your comment data is processed.