Researchers at Meta AI and the University of Texas developed AI that can “read” thoughts — converting brain activity into text in real time. The technology promises to transform the lives of paralyzed individuals, but raises enormous privacy questions.
📖 Read more: Meta AI: Accounts After Your Death
🧠 How It Works
The technology uses non-invasive fMRI (functional magnetic resonance imaging) combined with a large language model (LLM). The brain thinks a sentence, fMRI records hemodynamic patterns, and AI translates them to text. Accuracy now reaches 82% for simple sentences — over 70% improvement compared to two years ago.
The system doesn't “hear” exact words — it recognizes semantic patterns. If you think “I want water,” it might produce “user expresses thirst” or “needs water.” The meaning is conveyed, not word-by-word translation.
📖 Read more: AI Fashion Sizing: No More Wrong Sizes
♿ Healthcare Applications
The most significant application concerns people who cannot speak. Patients with ALS, stroke, or spinal cord injury could communicate through thought. Meta AI's team is already piloting in hospitals: locked-in syndrome patients managed to “say” basic phrases through fMRI.
BrainGate, Musk's Neuralink, and Synchron are developing similar systems but with implants. Meta's approach is non-invasive — no surgery needed. This is a huge advantage for broad application, though it limits accuracy compared to implants.
📖 Read more: Apple Touchscreen MacBook: What We Know
🔒 The Ethical Questions
The ability to “read thoughts” raises concerns straight out of Black Mirror. If the technology becomes portable (e.g., a special headband), can someone “read” your thoughts without consent? Can employers request brain scans? Can police use such technology?
Researchers emphasize the system requires extensive per-person training (15+ hours of fMRI) and active cooperation — it cannot “steal” thoughts without the person's willingness. But the technology evolves fast and guardrails must be established now.
📖 Read more: ASML Beyond EUV: The Future of Microchips
💡 Critical: UNESCO is already calling for “neuro-rights” — cognitive liberty (right to thought), mental privacy (privacy of thoughts), and mental integrity (brain protection). Chile is the first country to constitutionally enshrine these.
🔮 Where We're Headed
Meta targets a portable device (non-fMRI-based) by 2030, possibly based on EEG or fNIRS. If achieved, “telepathic” communication will no longer be science fiction. The question isn't whether it will happen — but whether society will be ready.
