AI Is Getting Closer to Reading Minds

For decades, the idea of uploaded intelligence has been the stuff of science fiction. From Pantheon to The Matrix, we’ve imagined worlds where human consciousness could be digitized and stored. But with the latest advancements in brain-computer interfaces (BCIs), we may be closer to this reality than we ever thought possible.

The Breakthrough: AI That Reads Your Thoughts

A recent study from Meta AI and leading neuroscientists (Lévy et al., 2025) has introduced Brain2Qwerty, an AI-powered system that translates brain activity into text using non-invasive methods. Unlike Neuralink and other brain-implant technologies, this system works using Magnetoencephalography (MEG) and Electroencephalography (EEG) to capture neural activity and decode it into readable text.

How It Works

The process is both intricate and fascinating:

• Capturing Brain Signals: Participants in the study were asked to type sentences while their brain activity was monitored via Magnetoencephalography (MEG) and Electroencephalography (EEG).
• AI-Powered Decoding: The Brain2Qwerty system translates these neural signals into readable text through a sophisticated three-stage AI model:
  1. A Convolutional Module that processes short bursts of neural data.
  2. A Transformer Module that refines sentence-level understanding.
  3. A Language Model that enhances accuracy and corrects errors.
• High Accuracy Levels: Using MEG, Brain2Qwerty achieves a 32% character error rate (CER), and in the best cases, the error rate drops to an impressive 19%.

Why This Changes Everything

These advancements bring us closer to a world where:

• People with disabilities can communicate more effectively using thought-driven text generation.
• Human-computer interaction evolves beyond typing and speech, enabling direct brain-to-device communication.
• AI assistants anticipate human needs, responding intuitively without spoken commands.
• Cognitive augmentation becomes a reality, opening new possibilities for memory storage and digital consciousness.

The Road to Digitizing Thought

While this research doesn’t yet achieve full brain uploading, it lays essential groundwork for future AI systems capable of:

1. Translating neural signals into fully formed thoughts and sentences in real time.
2. Storing and retrieving human experiences, memories, and cognitive patterns.
3. Developing mind-controlled computing systems that replace traditional interfaces.

The Big Questions: Ethics and Challenges

The ability to decode human thoughts raises profound questions:

• Privacy and Security: Who controls and protects the data extracted from brain signals?
• Ethical Boundaries: Should governments or corporations have access to our mental processes?
• Surveillance Risks: Could this technology be exploited for unauthorized monitoring?

What’s Next? The Future of AI-Human Integration

We are on the brink of a new era in AI and neuroscience. As non-invasive BCIs continue to advance, we may soon see a world where direct thought-to-text communication is not just possible, but practical.

What does this mean for society? Will AI serve as an extension of human intelligence or challenge our autonomy?of a new era in AI and neuroscience. As non-invasive BCIs continue to improve, we may soon see applications that allow direct thought-to-text communication, potentially paving the way for digitized intelligence.

What Do You Think?

Are you excited or concerned about AI reading thoughts? Could this technology be the first step toward true human-AI symbiosis?

Let’s discuss in the comments.

Reference:
Lévy, J., Zhang, M. (Lucy), Pinet, S., Rapin, J., Banville, H., d’Ascoli, S., & King, J-R. (2025). Brain-to-Text Decoding: A Non-invasive Approach via Typing. Meta AI Research.