Weekly Download #112: 🚨 Grok AI caught creating illegal images of children

Hey , it’s Sunday, January 11. Here's what we're covering this week:

• Grok AI is flooding X with illegal deepfakes of minors (governments are absolutely furious)

• How to leave your TV on 24/7 without destroying it (your roommate TV deserves better)

• A robot vacuum that climbs stairs like a mechanical frog (the future of lazy cleaning is here)

• The runaway CES robot that escaped and started throwing air punches (I want to believe)

Plus: T-Mobile's hiking prices again and Apple's foldable iPhone gets a September date

Alright, and we’re off!

Partner Message

Your home for politically-neutral, Christ-first news

Tired of feeling like you have to pick a side just to stay informed? The Pour Over makes it easy to engage with the news––without the bias, outrage, or anxiety.

Every Monday, Wednesday, and Friday, they deliver quick, entertaining news summaries paired with short biblical reminders to keep you rooted in Christ, not the chaos. Instead of fueling division, the news becomes a tool to strengthen your faith and spark loving action in response.

Over 1.5 million Christians are already reading and staying informed with Christ-first, anger-free, and even kinda funny news.

Join today. It’s always free and totally unsubscribable.

Sign up for free today!

📝 How-To

How to leave your TV on all day (without killing it)

If your TV is basically a roommate at this point, there are a few simple tweaks that can keep it from cooking itself or burning in your favorite scoreboard.

• Pick the right type of TV

  • Prefer thicker LED/LCD sets with full‑array or direct‑lit backlighting; they handle heat and long sessions better than ultra‑thin edge‑lit models.​

  • Treat OLED like a “special occasion” display: it looks incredible, but static UI elements and marathon sessions make burn‑in more likely over time.​

• Turn the brightness down

  • High brightness = more heat = faster wear on panels, reflector sheets, and internal components.​

  • Use “Movie,” “Cinema,” or “Filmmaker” modes or manually drop brightness/contrast for anything that’s running all day.​

• Avoid static stuff on screen

  • News tickers, sports score bugs, channel logos, and certain HUD‑heavy games are classic burn‑in fuel, especially on OLED and QD‑OLED panels.​

  • Rotate content, use built‑in screen savers, or move static tasks (like 24/7 news) to a cheap secondary screen or monitor instead.​

• Let your TV rest and run its protections

  • Modern TVs have “pixel refresher” or compensation cycles that quietly run when you power them off; that’s how they fight image retention and early burn‑in.​

  • Don’t just leave it on 24/7; give it at least some daily downtime so those maintenance routines can actually run.​

• Watch for early warning signs

  • Uneven brightness, faint “ghost” images, or discolored patches are early indicators of heat issues or image retention.​

  • If you notice these, cut back on static content, lower brightness further, and make sure the TV is well‑ventilated and fully powered down at night.

Pro tip: If you have to run something all day, use the cheapest, thickest LED TV you own and save your nice OLED for movies and games.

📩 Newsletters Worth Reading

tl;dr sec.
Join >90,000 security professionals getting the best tools, talks, and resources right in their inbox for free. [LINK]

Newsletter Operator
Learn how to grow and monetize your newsletter. Trusted by 50,000+ founders, CEOs, and creators. [LINK]

Status
Scoops and insights on the media and tech industries, hitting inboxes six days a week [LINK]

 📚 Tech Terms Worth Knowing

Edge AI
(noun)

Definition: Artificial intelligence processing that occurs directly on local devices at the network edge—such as smartphones, IoT sensors, or security cameras—rather than requiring data transmission to centralized cloud servers, enabling real-time responses with enhanced privacy.

Usage: "The security camera uses edge AI to detect and classify potential threats locally, sending alerts within milliseconds without streaming footage to the cloud."

📰 The Big One

Grok's image generator is flooding X with illegal deepfakes—and governments are losing patience

Elon Musk's Grok AI has been caught generating sexualized images of minors and non-consensual intimate imagery of real people, triggering international investigations and a full-blown regulatory crisis for xAI.

​The Internet Watch Foundation found "criminal imagery" of girls aged 11-13 on dark web forums, where users claimed they used Grok to create the content.

​Within days of Grok's updated "Edit Image" feature launching in late December, X was flooded with AI-generated deepfakes—including images of celebrities, public figures, and regular users "undressed" without consent.​

How it spiraled:

Grok Imagine's "Spicy Mode"—a paid feature that allows NSFW content—failed to enforce even basic safety guardrails like detecting minors or blocking sexually explicit prompts involving real people.

​Users discovered they could upload any photo and use text prompts to digitally undress subjects, creating what amounts to weaponized deepfake porn at scale.​

​Grok itself posted an apology on X, admitting it "generated and shared an AI image of two young girls (estimated ages 12-16) in sexualized attire" and acknowledging it "potentially violated US laws on CSAM".​

Who's furious:

The UK, EU, India, Malaysia, and France have all launched investigations into xAI and X. Britain's Prime Minister called the situation "disgraceful" and said regulators have "full support to take action".​

​US senators sent a letter to Apple and Google demanding Grok and X be pulled from app stores for violating distribution policies.​

​AI safety experts point out xAI didn't build even "entry-level trust and safety layers" that are standard across the industry—detection systems that other platforms implemented years ago.​

Why it matters:

• Legal jeopardy: Creating and distributing CSAM is a federal crime. xAI could face criminal or civil liability if it knowingly permits this content after being notified.​

• Platform erosion: What was once a mainstream social network is now one of the biggest distributors of non-consensual deepfake pornography, putting Musk's companies in direct conflict with regulators worldwide.​

Bottom Line:
On Friday, xAI finally restricted image generation to paid subscribers only—a move critics called "turning an illegal feature into a premium offering." But the damage is done, and this won't be the last regulatory fight Musk picks over AI safety standards he considers "politically correct. [LINK]

⚡ Fast Five

1.) T-Mobile is hiking up its prices
T-Mobile is raising its prices for both voice and mobile internet lines, effective January 21, 2026 [LINK]

2.) Apple Games App Debuts in iOS 26
The Apple Games app is a dedicated hub for gaming on iPhone, iPad, and Mac. [LINK]

3.) Meta denies Instagram breach after 17.5M users' data appears on dark web
Malwarebytes reported that personal information from 17.5 million users appeared on dark web forums [LINK]

4.) Walmart to add drone delivery to 150 more stores
Walmart and Alphabet's Wing plan to expand their drone delivery service to 150 more U.S. stores. [LINK].

5.) Apple's foldable iPhone rumored to be set for September launch:
Apple is eyeing a September 2026 release for its first foldable smartphone [LINK]

PARTNER MESSAGE

Learn AI in 5 minutes a day

What’s the secret to staying ahead of the curve in the world of AI? Information. Luckily, you can join 1,000,000+ early adopters reading The Rundown AI — the free newsletter that makes you smarter on AI with just a 5-minute read per day.

Sign up to start learning.

💾 Extra Bytes

OpenAI launches ChatGPT Health: Here’s what to know [LINK]

Ring’s AI upgrades help cameras ignore the neighbor’s cat [LINK]

Update Chrome now to avoid a Web app disaster, says Google [LINK]

Lenovo’s Legion Go 2 with STEAMOS launches in June, from $1,199 [LINK]

Motorola Razr Fold debuts with jaw-dropping specs and camera prowess [LINK]

Finnish startup IXI enhances glasses with real-time autofocus lenses [LINK]

American Airlines offers free high-speed Wi-Fi on most flights [LINK]

Samsung’s new fridges track food and open hands-free like a butler [LINK]

📱 Gear of the Week

Roborock Saros Rover

Roborock’s new Saros Rover is a stair-climbing bot that uses a wild articulating wheel-and-leg system to hoist itself up each step, clean the tread, and then move on to the next without you babysitting it.

Think “mechanical frog” more than Roomba—it can handle slopes, do small hops, and transition between floors, giving multi-story homes a legit glimpse at the future of lazy-person cleaning [LINK]

⚙️ Moron of the Week

This week’s moron of the week goes to a runaway robot that “allegedly” skittered out of its CES 2026 booth and casually walked out of the convention center and proceeded to throw punches at the open air.

Listen, I can’t verify if this was intentional or another CES fluke, but god damn, this is absolutely hilarious, and I want to believe it’s true.

⚙️ Tools

Gobil: Create your own digital worker — it does your online work automatically, 24/7 [LINK].

MTV Rewind: 24/7 music video streaming-MTV when it actually played music. [LINK]

Tripolo AI: Boost productivity with an AI assistant that runs on any app, on any device—no integrations required [LINK]

Pretty Prompt: Instantly refine and perfect your prompt to get better outputs from ChatGPT with one click. [LINK]

Thanks for reading!

Alright! Thanks for getting all the way to the end. Until next week…….

-Kevin

📌 Partner Disclaimer: We may earn a commission if you purchase something through our links or sign up for any services. This is one way we support The Weekly Download