Skip to content
noisebrief.

Today's tech noise. Briefly.

One email. Every morning.

Past brief

Monday, March 16

AI's Human Cost Emerges

The human cost of AI training has taken disturbing new forms as the industry's appetite for authentic data grows. Professional models now earn money making up to 100 video calls per day (opens in a new tab) to lend their faces to AI scams, creating a cottage industry around deception. Meanwhile, gig workers are being paid to film their daily chores to train robots, turning mundane household tasks into commodified training data. AI companies are even recruiting improv actors (opens in a new tab) to harvest their emotional skills, seeking to capture the nuances of human feeling for machine learning systems.

Big Tech's strategic shifts reveal cracks in the AI integration playbook that dominated 2023-2024. Microsoft (opens in a new tab) has abandoned plans to inject Copilot into a key part of Windows 11, signaling potential user resistance to ubiquitous AI features. ByteDance (opens in a new tab) paused the global launch of its Seedance 2.0 video generator, suggesting international regulatory pressures may be forcing content companies to reconsider their AI rollout strategies. These retreats come as Wall Street (opens in a new tab) begins betting on prediction markets, indicating that financial institutions are hedging their AI investments with alternative data-driven revenue streams.

A wave of biotech breakthroughs promises to reshape healthcare through direct neural and biological interfaces. Scientists have created the first artificial neuron (opens in a new tab) capable of communicating with the human brain, building on 2014's Medical Body Area Network approval that established wireless communication protocols for signals inside the human body. Separately, researchers have developed working hair follicles in laboratory conditions that could permanently cure baldness, while fitness trackers (opens in a new tab) are proving unexpectedly valuable as diagnostic tools for chronic illness management.

Sources · 66