We are in the midst of a mental privacy panic. Again. By now, as the legal academic Francis Shen has observed, they have a well-worn script. First, note we are on the verge of powerful mind reading technologies. Second, argue the state will use these in devious ways. Third, claim citizens will be powerless to stop this because the law is not ready. Finally, call for immediate action to prevent the government from reading our minds. If you can reference George Orwell or Philip K Dick, all the better.
Five years ago, such cynicism was justified. But things have changed.
We are no longer “on the verge” of mind reading technologies. We have them. We need not worry if they will be used in devious ways. They have. It is not only the government we need fear. It is private companies too. This became clear in the extended fallout from the Cambridge Analytica scandal, in which a firm harvested Facebook data to target users with personalised political ads based on their psychological profile.
Mind reading is nothing new: we are constantly trying to work out what other people are thinking. We scrutinise facial expressions, read body language, and listen to the tone of voices. We accept this as fair, assuming that most adults can keep certain information hidden. Unfortunately, machines can now overcome our mortal ability to keep our mental world private.
There are two ways to go beyond using the data you voluntarily share on a Facebook profile, and get into your head. Both are in their infancy, with studies in need of replication. Yet, as large sums of money pour into such work, progress may be rapid.
The first route in is brain reading. This measures brain activity and attempts to decode the thoughts it represents. It is easy to overstate what neuroscience can currently do, and we need neuromodesty. Nevertheless, there is enough to warrant concern.
A 2001 study found data from brain scans could predict with a high degree of accuracy what type of object (e.g., face, house, cat) someone was looking at. You think your dreams are safe? Think again. Research has found that brain data obtained when people were dreaming predicted with 60 per cent accuracy what categories of objects featured in their dreams. Not only can neuroscience predict what people are seeing, it can reconstruct it. A 2008 study took participant’s neural activity and used it to reconstruct what they were actually seeing.
Other research is trying to work out what words we are saying to ourselves silently in our heads. Approaches that measure neural activity using non-invasive brain scans or by placing electrodes on the brain itself are showing success here. Nasa is also working on this problem, using sensors to on the throat and chin to detect and decode inner speech. They hope this will allow astronauts to operate machines in space with their minds alone.
Intentions can be decoded too. A 2007 study asked people to decide whether they wished to either add or subtract two numbers. Analysis of the activity in their prefrontal cortex (involved in planning) allowed researchers to predict which option the person had chosen with a 71 per cent accuracy. Knowing someone’s brain activity can even allow us to predict what someone will do (for very basic tasks) before they themselves know.
Even political stances can be detected. A 2011 study reported the size of some brain regions predicted participant’s levels of liberalism/conservatism. One of these regions was the brain’s fear centre – the amygdala. A later study found brain activity during financial risk taking allowed the correct prediction of the political affiliation (Republican or Democrat) of 83 per cent of participants. Again, the amygdala was one of these areas.
Much of this work is adolescent fumbling. Accurate predictions about people’s thoughts are only possible under highly constrained experimental conditions. Due to individual variability, the brain activity underlying my thoughts may be very different to yours. This poses problems for a universal brain reading machine. That said, the brain activity underlying a particular memory in one person, can be used to predict if another person is experiencing that memory.
Big tech companies are throwing money at this. Facebook has announced plans for a Brain Computer Interface. This aims to decode users’ thoughts and transmit them to Facebook. Microsoft has patented brain reading technology. Elon Musk’s startup, Neuralink, is also trying to develop brain-computer interfaces.
Governments are also spending. The US National Institutes of Health believes such work will help people paralysed by spinal injury to communicate. However, this work also has military applications, such as improving soldiers’ performance. As such, it has attracted funding from the the United States Army, the United States Air Force and the Intelligence Advanced Research Projects Activity. There may also be other national security interests in this research.
The second way to get into your head is behaviour reading. This gathers data on your observable behaviour, such as facial expressions, actions, possessions, and “likes” on Facebook. This is then used to infer what is happening in your inner world.
Facial analysis technology is increasingly adept at detecting our emotions and much more beside. Deep learning neural networks are better at detecting people’s sexuality from their faces than people are, according to a controversial study published earlier this year. The researchers stated that these “findings expose a threat to the privacy and safety of gay men and women.”
People’s musical preferences predict aspects of their personality. The words we use on Facebook are markers of personality. What you “like” on Facebook is highly informative. It was a shock to many people when research reported that detailed personal information could be predicted with a high level of accuracy from what pages on Facebook people had “liked”.
The law saw this coming. “Advances in the psychic and related sciences,” suggested US Supreme Court Justice Louis Brandies in 1928, “may bring means of exploring unexpressed beliefs, thoughts and emotions.” On paper, at least, human rights law polices the borders of our mind. Besides the right to privacy, most of us assume we also have an absolute right to freedom of thought, which includes the right to keep our thoughts private. This right desperately needs developing in relation to technological advances.
The right to freedom of thought also prevents others either manipulating our thoughts or punishing us for them. Again, we will need to consider this law in relation to contemporary developments. The proposed social credit system in China could lead to punishment for thoughtcrime. Knowing our thoughts also opens us up to both more effective commercial and political manipulation.
We are not powerless to stop non-consensual incursions into our inner world. But, we need to establish how law is going to operate in an era in which machines make us increasingly transparent. This is a matter of urgency. “Carefully guard your thoughts”, runs one translation of Proverbs 4:23, “because they are the source of true life”.