AI News Roundup: 2026-03-24

Here is your weekly roundup of the most important AI stories, what they mean, and why they matter to regular people. This week's topics include:

• Thousands Selling Their Identities to Train AI
• The AI Identity Marketplace: Trading Biometrics for Grocery Money
• AI Detecting 20% More Breast Cancers in Major Studies
• Cognizant: 93% of Jobs Face AI Disruption, Six Years Early
• Val Kilmer to Appear Posthumously in Film Using AI
• Anthropic's Consumer Surge After Privacy Stance
• AI Computing's Big Shift: Inference Overtakes Training

Thousands Selling Their Identities to Train AI

2026-03-24

DoorDash just launched a program redirecting 8 million delivery drivers to a new kind of work: filming themselves loading dishwashers and folding laundry to train robots. Uber did the same thing in October. The Pentagon is discussing letting AI companies train on classified military data.

The data gold rush is global. A student in India earns $100 a month recording city noise. A welder in Chicago sold his private family chats for $0.50 per minute. A man in Cape Town made half a week's groceries filming his feet while walking.

For people in countries with devalued currencies or high unemployment, earning US dollars feels like a lifeline. And it is, right now. But Oxford researchers warn this work is "precarious, non-progressive, and effectively a dead end." You grant irrevocable licenses. Your 20-minute voice recording powers a customer service bot for years. You never see another cent.

Here's the contradiction: individuals getting paid today for data that tech companies have been taking for free for decades. That feels like progress. But the same data marketplace that gives people income now may lock them out of jobs later.

Which matters more in the long run: getting paid for your data today, or protecting against a future where that data replaces your livelihood?

#DigitalRenAIssance #AI

Read the full story on theguardian.com

The AI Identity Marketplace: Trading Biometrics for Grocery Money

2026-03-23

A 27-year-old in Cape Town recorded videos of his feet walking on pavement and earned $14. A student in India made $100 a month letting an app record ambient city noise through his phone. An 18-year-old in Chicago sold his private text conversations with friends and family for $0.50 per minute.

They're part of a new global data marketplace. As AI companies run out of high-quality training data, thousands of people are micro-licensing their voices, faces, and personal conversations to platforms like Kled AI, Silencio, and Neon Mobile. The pay is real. So are the consequences most of them don't yet understand.

When you sell your likeness on these platforms, you're granting a worldwide, exclusive, irrevocable license. Adam Coy, an actor from New York, sold his face and voice for $1,000 with protections against political misuse. Not long after, his friends forwarded him videos of his AI replica promoting unproven medical supplements to pregnant women as a "vagina doctor." Millions of views. Not actually him.

Meanwhile, identity fraud hit $47 billion in 2024. Deepfake usage in fraud surged 58% year over year. Voice cloning now takes just 30 seconds of clean audio scraped from a LinkedIn video or conference call. A finance worker transferred $25 million after a video call with what appeared to be their CFO and colleagues. All deepfakes.

The calculation is simple for people selling their data: tech companies already capture so much of our information, why not get paid? But here's the paradox. They're fueling the same industry that's training the tools used to exploit them.

Is this a fair exchange or are we trading pennies today for risks we can't yet see?

#AI #DigitalRenAIssance

Read the full story on theguardian.com

AI Detecting 20% More Breast Cancers in Major Studies

2026-03-20

AI just caught 20% more breast cancers than human doctors working alone.

Three major studies across Sweden, Spain, and the UK all show the same pattern. AI-supported mammogram screening detected 15-25% more cancers, especially early-stage ones, while reducing false positives and unnecessary biopsies. In the largest NHS study ever, 175,000 women got their scans analyzed by AI in 17.7 minutes instead of waiting 2.08 days for the first human reader.

This isn't replacing radiologists. It's giving them a partner that never blinks, never gets tired, and catches the subtle signs human eyes miss. And it's freeing them to focus on the high-risk cases where their expertise matters most.

Here's what gets me. This technology is already widely used across Europe. The NHS does 2 million mammograms per year with a 29% radiologist shortage projected to hit 39% by 2029. AI isn't just improving detection, it's solving a workforce crisis while saving lives.

Meanwhile, in the US, we're still running trials. Yes, we need to address AI bias (Black women see 50% more false positives). Yes, we need radiologists in the driver's seat. But how many lives could we save while we perfect it?

Is caution keeping us safe, or is it costing lives we could be saving right now?

#DigitalRenAIssance #HealthTech

Read the full story on breastcancer.org

Cognizant: 93% of Jobs Face AI Disruption, Six Years Early

2026-03-19

Cognizant just updated their AI workforce forecast. The results came six years early and hit harder than expected.

The numbers: 93% of jobs now face some AI disruption. 30% face existential threats, up 15 percentage points from 2023. The kicker? It's not just white-collar work. Construction, transportation, healthcare. AI is augmenting tasks once considered purely manual.

At Cisco, I led teams of 500+ employees. Today, companies like Block achieve similar output with half the workforce. This isn't a future trend. It's happening now.

Here's the reality check. Matt Sigelman from the Burning Glass Institute said it best: "People who have been in a job for decades may no longer be qualified for the job that has defined their careers." That's terrifying if you're coasting. Liberating if you're learning.

What skills are you building today that make you more valuable alongside AI, not replaceable by it?

#DigitalRenAIssance #AIForEveryone

Read the full story on fortune.com

Val Kilmer to Appear Posthumously in Film Using AI

2026-03-19

Val Kilmer will appear in a new film. He passed away in April 2025.

The movie is called "As Deep as the Grave," and Kilmer was cast as Father Fintan back in 2020 but couldn't shoot due to the throat cancer he battled for over a decade. Rather than recast the role, director Coerte Voorhees used generative AI to digitally create Kilmer's performance. It's believed to be the first time a posthumous film role has been built entirely with AI rather than CGI and body doubles.

Here's what makes this different from the usual AI fear story. Kilmer's children, Mercedes and Jack, gave their full blessing. Mercedes said her father "always looked at emerging technologies with optimism as a tool to expand the possibilities of storytelling." AI previously recreated his voice for Top Gun: Maverick. This was the next step he wanted to take.

The timing matters. SAG-AFTRA just failed to reach a deal with studios, and AI usage in film remains the biggest sticking point. California law AB 1836 now protects digital likenesses after death. Consent and family approval aren't optional anymore. They're the law.

When the actor says yes, the family says yes, and the law protects the decision, is this the future of honoring an artist's legacy or the beginning of something we can't take back?

#AI #DigitalRenAIssance

Read the full story on nbcnews.com

Anthropic's Consumer Surge After Privacy Stance

2026-03-18

Anthropic drew two red lines with the Pentagon: no fully autonomous weapons, no mass surveillance of Americans. Defense Secretary Hegseth's response was blunt: "We will not employ AI models that won't allow you to fight wars."

The company stuck to its principles anyway. The result? Claude hit #1 on the App Store in 16 countries. Daily active users tripled since January to 11.3 million. More than 1 million people are signing up every day.

Here's the part that matters. Nearly 150 retired federal and state judges filed an amicus brief supporting Anthropic. Chalk messages appeared outside their headquarters: "You give us courage." Meanwhile, ChatGPT's uninstalls surged as OpenAI's opportunistic Pentagon deal drew backlash.

For thirty years, I've watched tech companies face this choice. Most pick short-term revenue over long-term trust. The rare ones that choose principles often get punished for it. Not this time.

When companies stand for something real, consumers notice. When they prioritize human safety over government contracts, people vote with their downloads.

What matters more when you choose AI tools: the capabilities or the ethics behind them?

#AI #DigitalRenAIssance

Read the full story on techcrunch.com

AI Computing's Big Shift: Inference Overtakes Training

2026-03-18

For five years, AI companies chased a simple equation: build bigger models with more training compute, get smarter AI. That equation just flipped.

Deloitte's latest predictions reveal something counterintuitive. Two thirds of AI computing in 2026 will be inference, using trained models to answer questions, not training. The surprise? Inference now demands MORE compute power, not less, because of test-time scaling. The same reasoning breakthroughs that make AI feel smarter also make it more power-hungry.

The numbers are staggering. AI data center spending hits four hundred to four hundred fifty billion dollars this year and could reach one trillion by 2028. Microsoft added one gigawatt of data center capacity in a single quarter. That's enough power for a mid-sized city.

This isn't about training fewer models. It's about billions of people using AI every single day, and each query requiring serious computational firepower to deliver that instant, thoughtful response. The infrastructure we're building today determines who can afford to serve AI at scale tomorrow.

What happens when AI inference costs more to run than training the models in the first place?

#DigitalRenAIssance #AI

Read the full story on deloitte.com

Want to get these updates as they happen? Follow me on LinkedIn or X (Twitter).