Article By Frank Bergman
A group of researchers has sent shockwaves through the scientific community and triggered panic among experts by unveiling new technology that uses artificial intelligence (AI) to read humans’ minds.
Scientists have unveiled an AI-powered system that reconstructs what people see by analyzing their brain scans.
The breakthrough has raised alarming concerns about coercive interrogation, predictive policing, and intrusions into the privacy of the human brain.
Professor Michal Irani of the Weizmann Institute of Science developed Brain-IT, an AI model that interprets functional magnetic resonance imaging scans to recreate images viewed by participants.
Researchers are now looking beyond still images toward hearing, video, and potentially dreams.
However, many experts are sounding the alarm about the scientific breakthrough.
The United Nations (UN) warns that emerging AI brain-scanning neurotechnology could allow governments to extract information from people’s minds and treat their thoughts as evidence.
UN Warns of Forced Information Extraction
UN Secretary-General António Guterres warned Thursday that neurotechnologies are advancing rapidly, raising the prospect of authorities using brain data in investigations, court proceedings, and prisons.
The scenarios include officers extracting information from a suspect’s brain or scanning people’s minds to predict who might commit a future crime.
Guterres warned that abuses could lead to arbitrary detention, the alteration of mental states, and profound violations of privacy.
“Neurotechnology, including new and emerging forms, raises profound human rights challenges when applied to the administration of justice,” Guterres wrote in the report.
“Its expansion beyond clinical settings into investigative, judicial and correctional contexts introduces significant risks, alongside potential benefits,” Guterres added.
The UN’s first formal warning comes as researchers demonstrate increasingly sophisticated methods for translating brain activity into information that machines can interpret.
AI Reconstructs Images From Brain Activity
Irani and her colleagues released their Brain-IT research in a paper titled “The New Science of Mind-Reading.”
The team included Roman Beliy, Amit Zalcher, Jonathan Kogman, and Navve Wasserman from Weizmann’s Computer Science and Applied Mathematics Department.
The study examined fMRI scans from eight participants who viewed thousands of images across 30 to 40 scanning sessions.
Participants viewed around 40 images during roughly 10-minute sessions while researchers scanned their brains six times.
The work produced approximately 73,000 image-fMRI pairs for the AI model to analyze.
Brain-IT reconstructs viewed images with detailed content and composition. It can also predict brain activity by analyzing what a person is looking at.
Irani said other models produce “impressive reconstructions” but frequently make basic mistakes involving composition and color.
“The new model we developed outperforms them in reconstructing both the content of the image and its details,” Irani said.
The team also sharply reduced the scanning time needed to adapt the technology to a new person.
“What’s more, while every other model requires dozens of hours of brain scans to learn to ‘read’ a new person, our model needs only one hour,” Irani said.
Researchers Aim to Decode Dreams
The current study used fixed images. Irani’s laboratory plans to test the methods on hearing and eventually video.
Moving images pose a technical challenge because visual information changes much faster than an fMRI machine can capture it.
“What remains especially challenging is decoding video — for example, during dreaming,” Irani said in a statement.
“Dozens of images change every second, while an fMRI scan takes about two seconds.
“If we overcome all these obstacles, it’s possible that in the future we may even be able to read dreams,” Irani said.
That ambition would extend the technology beyond reconstructing pictures shown to participants and into experiences occurring during sleep.
Warnings Over Coercion and Mental Privacy
Scientists have explored brain-reading technology for years, including its possible use in the justice system.
A 2017 paper published in the Journal of Law and the Biosciences discussed “brain-based mind reading” through neuroimaging techniques.
The paper examined potential applications involving lie detection and assessments of defendants, prisoners, and prospective jurors. It also identified the danger of unethical, coercive uses.
Some neuroscientists have similarly warned that tools such as Brain-IT could be abused to extract information from a person’s mind against their will.
Other researchers have highlighted possible therapeutic applications for neurological conditions, depression, and anxiety.
Princeton University psychology and neuroscience professor Kenneth Norman explained in a 2022 American Psychological Association discussion that improvements in fMRI signals and their analysis had helped make the technology possible.
Norman described how therapists could use neurofeedback to help patients recognize negative thought loops and learn to control those patterns.
But the UN warning reaches far beyond treatment: the same expanding ability to interpret brain activity could place deeply private information within reach of investigators, courts, and prison authorities.
Brain-IT has demonstrated image reconstruction from brain scans.
Its developers’ ambitions now include dreams, while Guterres is warning about a future in which governments could use information extracted from the brain against the person it belongs to.
Be the first to comment