In partnership with

Welcome Automaters, 👋

So… remember when Meta promised their Ray-Ban smart glasses were totally private and "controlled by you"? Yeah. About that.

A bombshell investigation by Swedish newspapers recently revealed that workers at a Kenya-based subcontractor were literally watching footage recorded by Meta's AI glasses. And we’re not just talking about boring grocery runs here. We are talking about nudity, sex, and people using the bathroom. Basically, it’s a reality show that absolutely nobody signed up for. 😬

Now Meta is getting sued. Two plaintiffs (Gina Bartone of New Jersey and Mateo Canu of California)  filed a class action complaint through the Clarkson Law Firm (the same folks who've gone after Apple, Google, and OpenAI), alleging Meta violated privacy laws AND engaged in blatant false advertising

Why false advertising? Because Meta's own ads literally said things like "you're in control of your data and content" and "built for your privacy." The plaintiffs believed those promises, and shockingly, nobody mentioned the part about overseas workers reviewing intimate moments. Wild, right?

Meta’s defense is that they take steps to “filter this data to protect people’s privacy and to help prevent identifying information from being reviewed.” But according to sources who spoke to the BBC , that filtering of data? Not always working. The U.K.'s Information Commissioner's Office has already launched its own investigation. And when the privacy watchdog starts writing letters, you know the situation is getting serious.

To be fair, Meta does mention human review in its fine print — buried deep in its terms of service. But come on. Seven million people bought these glasses in 2025 alone. How many of them actually read the terms? (Spoiler: approximately zero of them.)

The Big Picture:

This isn't just a Meta problem: it’s a wake-up call for the entire AI wearables industry.

As AI gets baked into everything we wear, eat, and breathe, the gap between "sounds cool in an ad" and "here is what is actually happening to your data" is becoming a canyon. Between AI pendants, always-on microphones, and smart glasses, we are building a world of luxury surveillance and the legal system is just starting to catch up, but for now, the lesson is clear: 

If it has a camera and "AI" in the name, someone might be watching.

Also: Here’s a little something from us for the weekend:

Go watch it and let us know what you think!

Here's what we have for you today

🤖 OpenAI GPT-5.4 Launch: Everything You Need to Know About the New ChatGPT Model

OpenAI just released their most powerful model yet, and the ink on GPT-5.3 instant is barely dry.

Just two days after OpenAI gave ChatGPT its highly-memed "cringe-ectomy" with GPT-5.3 Instant (the update that literally deleted phrases like "Stop. Take a breath." from ChatGPT's vocabulary), now, the company dropped GPT-5.4 on Thursday, March 5th. I mean two major model drops in days. OpenAI isn't walking: they're sprinting.

So What's Actually New?

GPT-5.4 is tagged as OpenAI's "most capable and efficient frontier model for professional work." And it comes in three distinct versions:

  • GPT-5.4 Standard: The balanced workhorse.

  • GPT-5.4 Thinking: The reasoning specialist (perfect for deep strategy or complex logic).

  • GPT-5.4 Pro: The high-performance beast for maximum accuracy.

Think of it like ordering coffee: regular, extra shot, or "I have a board presentation in 20 minutes and I need to build a financial model, a legal brief, and a slide deck right now." That last one is GPT-5.4 Pro.

For the tech-heads reading this: the API version now supports context windows as large as 1 million tokens. That’s the largest available from OpenAI to date. The company also introduced a "Tool Search" system for developers that reduces token consumption by nearly half.

The Benchmarks are honestly a little scary to:

  • In Professional Knowledge: It Scored 83% on the GDPval test (a jump from 70% in GPT-5.2), according to OpenAI, the model now matches or exceeds industry professionals across 44 occupations.

  • Computer Use: It set record scores on OSWorld-Verified (75%) and WebArena-Verified, officially surpassing the human baseline for navigating desktops.

  • Accuracy: OpenAI says 5.4 is 33% less likely to make false claims and 18% less likely to have errors compared to GPT-5.2.

  • Safety: OpenAI also rolled out a new safety test that checks how its models handle chain-of-thought reasoning. The result? The Thinking version of GPT-5.4 is way less likely to pull any sneaky moves. According to OpenAI’s evaluation, the model doesn’t seem able to hide its reasoning, which means monitoring its chain-of-thought still works as a pretty solid safety check.

But let's rewind 48 hours. On March 3rd, OpenAI rolled out GPT-5.3 Instant. Its whole vibe was fixing ChatGPT's deeply annoying habit of being overbearing.

We’re talking about those unprompted phrases like "You're not broken" or "I hear you" that made users feel like they were being handled by a guidance counselor instead of an AI tool. OpenAI heard the feedback loud and clear: they cut the "cringe," reduced hallucination rates by 26.8%, and made the bot sound like a peer again.

The Big Picture: 

OpenAI's rapid-fire release cadence isn't just a flex. It signals a fundamental shift. We're moving from "AI as a toy" to "AI as essential infrastructure."

With native computer-use capabilities and the ability to handle long-horizon deliverables (think investment banking pitch decks and full legal reviews), GPT-5.4 is no longer just a fancy search engine. It’s gunning for your lawyer, your financial analyst, and your slide deck designer simultaneously.

Smart starts here.

You don't have to read everything — just the right thing. 1440's daily newsletter distills the day's biggest stories from 100+ sources into one quick, 5-minute read. It's the fastest way to stay sharp, sound informed, and actually understand what's happening in the world. Join 4.5 million readers who start their day the smart way.

🧱 Around The AI Block

🤖 AI Workout Of The Day: How to Hack Your Style in 2026

AI Generated

Imagine if you could snap your fingers and your 2 AM notebook doodles turned into professional photos on actual models. Sounds like a scene from a Disney movie, right? Well, that is basically the reality of fashion in 2026. Except instead of magic wands, designers are using AI.

The fashion world is going through its biggest glow-up since the invention of the sewing machine. And the numbers? They are absolutely spicy. The AI fashion market was sitting at about $2.92 billion in 2025, but it is currently on a rocket ship headed for $89 billion by 2035. We are witnessing a 40% annual growth rate, which is the industry equivalent of going from a local lemonade stand to a global conglomerate in record time.

Here’s how AI is helping

  1. The Sketch-to-Runway (In seconds) Pipeline: Tools like FASHN, NewArc.ai, and Fashion Diffusion are the MVPs here. You upload a rough sketch and the AI generates a studio-quality photo of that outfit on a model of your choice. Small brand owners are obsessed because it lets them create professional lookbooks without the $10,000 price tag of a traditional shoot.

  2. Virtual Fit-Checks: Ever hit "buy" and hoped for the best? Those days are over. Giants like Zalando and Google have perfected virtual try-on tech. You upload a selfie and the AI maps the garment to your body, accurately showing how silk flows versus how denim stays stiff. Reports show this tech is already boosting conversion rates by 25% because people finally know if that dress will actually fit.

  3.  Predict What's Going to Be Cool Before It's Cool: This is where things get scarily smart. Zara, H&M, and G-Star Raw are using AI to scan millions of TikToks, Instagram posts, and celebrity "Outfits of the Day." They use Natural Language Processing (NLP) to "listen" to what people are saying about minimalist styles or "boho-chic" and then design collections to match in real-time. Zara can literally get a trend from a social media post to a store shelf in just seven days.

Other Heavy-Duty Tools:

  • CLO 3D & Style3D AI: These aren't just for "pretty pictures." They simulate the physics of fabric. Want to see how a heavy wool coat moves in the wind? These tools show you perfectly, cutting the need for physical samples by 70% to 90%.

  • FashionINSTA: This is the new kid on the block that focuses on "Pattern Intelligence." While other tools just make a cool image, this one converts your sketch into a production-ready sewing pattern in about 10 minutes.

  • VModel.ai & Botika: These are the "digital agencies" of the future. They let brands change the ethnicity, age, or pose of a model in an instant. It makes diversity representation easy and affordable rather than an afterthought.

  • Heuritech: This is basically a "Shazam for Fashion" that predicts trends with 90% accuracy.

Here's how it works in practice: A designer can sketch a jacket in the morning, use AI to create 20 different versions by lunchtime (changing colors, fabrics, patterns), then test which ones people like best by the afternoon—all without making a single physical sample.

The Catch: It is not all sunshine and sequins. There is a learning curve for older brands, and some designers worry about "technical soul." But the consensus for 2026 is clear: AI isn't replacing the designer. It is just giving the artist a much faster brush.

The Bottom Line: Whether you’re a global CEO, a teenager with a sewing machine, or a shopper trying not to regret their purchase, the barrier to entry has never been lower.

💡 Prompt to try:

For Use when you have a hand-drawn or digital sketch:

Convert this fashion/design sketch into a highly realistic, photorealistic image. Preserve the original design proportions, silhouettes, and details.

 -Style: ultra-realistic studio photography  
 -Materials: realistic fabric texture with natural folds and shadows  
 -Lighting: soft professional lighting, high dynamic range  
 -Camera: 85mm lens, shallow depth of field  
 -Quality: high resolution, sharp focus, true-to-life colors  

Do not stylize. Do not cartoonize. Make it look like a real product photo.

For a model image or yourself:

Place the clothing from the reference image onto the person in this photo. Ensure realistic fit, fabric drape, and body proportions.

Maintain natural skin texture, realistic lighting, and correct shadows. The clothing should follow the body’s posture and movement naturally.

 -Style: photorealistic fashion photography  
 -Lighting: natural daylight  
 -No distortion, no exaggerated body features, no unrealistic smoothing.

Clean Mannequin / Model Swap Prompt

Apply this outfit to a neutral fashion model.

 -Pose: relaxed, natural stance  
 -Background: minimal studio backdrop  
 -Focus on accurate tailoring, stitching details, and fabric texture  
 -Photorealistic, editorial fashion shoot quality

Is this your AI Workout of the Week (WoW)? Cast your vote!

Login or Subscribe to participate

That's all we've got for you today.

Did you like today's content? We'd love to hear from you! Please share your thoughts on our content below👇

What'd you think of today's email?

Login or Subscribe to participate

Your feedback means a lot to us and helps improve the quality of our newsletter.

Reply

Avatar

or to participate

More From The Automated