Bristol exhibition: public share feelings about technology and privacy

I recently held an interactive art exhibition in Bristol, UK where I explored people’s attitudes towards the use of their emotional and cognitive data. The findings were a clarion call to lawmakers.

There were two interactive artworks, one that responded to facial expression, the other to brain rhythms. When visitors stood before a camera and smiled, frowned or opened their mouth, the artwork, which was projected onto the wall in front of them, changed according to their expression.

In the second artwork ‘Bristol in Brain Rhythms’, visitors could put on an EEG (electroencephalogram) headband and watch the artwork, projected onto the wall, alter in line with their changing brain frequencies. Watch the video here: https://youtube.com/shorts/aWWs0b2cDM0

As part of the exhibition visitors were invited to share their feelings about the use of four different sorts of data: (1) voluntary hand gestures, (2) mood through facial expression, (3) brain rhythms, and (4) actual thoughts. They were asked how they’d feel if this information was used by an adaptive feature in their environment, such as digital wallpaper , that could sense and respond to them by changing design in real time. They were asked how they’d feel if the wall responded to them in different settings: (1) alone in a personal space (private bathroom), (2) with family/friends in their living room, and (3) in public (a restaurant). There was a fantastic level of participation with 27 visitors completing a questionnaire.

The findings were interesting. The fact that people were least comfy with their data being used by a public space, and most comfy in the private setting was not surprising.

Overall people were very comfy in all settings about wall décor responding to their deliberate hand gestures (alone or with friends / family a comfort score of 9 out of 10), and only marginally less so in the public space (8 out of 10). This isn’t surprising as deliberate gestures represent a type of data the individual willingly and knowingly gives.

People were very uncomfy when settings containing other people responded to their thoughts (with family / friends 3/10 and in a public place 2.5/10). Interestingly , they were not even particularly comfy with having their thoughts responded to in a private setting, suggesting a discomfort with thought reading irrespective of who sees it. The following comments were made:

“Your thoughts and brain are almost your only safe space.”

“My thoughts are the most private part of my human experience and they’re for no-one but me alone.”

On average, responders were neither particularly comfy nor uncomfy about having their mood (via facial expression) and brain rhythms captured, but there was a lot of variation in responses with some people nearly as happy to share their mood and brain rhythms as they were to make voluntary gestures, and others viewing this information as more private like thoughts.

“I’d be much more comfortable being able to see my thoughts and brain frequencies on my own than with others able to view them. Equally I wouldn’t like the reflection of my facial expressions in a vulnerable state.” — Exhibition visitor.

Gallery 2121: This part of the exhibition invited visitors to view art telling a story about the present time from the perspective of a future being. It aimed to encourage people to consider how our everyday lives may evolve.

People were slightly less comfortable with having their mood via facial expression captured than their brain rhythms. This might seem counterintuitive because brain activity is more internal than facial expression. Perhaps it is the meaning derived from the data which affects how sensitive people are about it; something that would explain the extreme sensitivity over thought data. If the question had concerned cognitive information derived from the brain rhythms responses would possibly have been more cautious.

One person commented:

“I hate the idea of people ‘getting into my head’. Thoughts are moments in time and often not fully processed views attitudes opinions. Many misunderstandings could occur. There’s also good & bad in all of us — revealing the uncensored bad would be scary.”

I also asked visitors if they thought their attitude might be different in 2050. These responses reflected a mix of people thinking they might become used to their data being captured, and those determined to uphold their conviction that this data should remain private:

“Yes, I feel I will be either desensitised to it or more understanding of it.”

“I think that if thought tracking technology were to be more normalised then I would maybe get desensitised to the idea.”

“I think I would still be troubled by my internal world being translated to the public in ways I can’t control.”

“I don’t believe they will as I find that a lot of these things will and should forever stay personal”

“I hope it never gets to the point that such practice becomes common.”

Technology has advanced rapidly making it easy for people to capture their mental and physiological states with wearables, but also for third parties to access that data. Moreover as tech advances, collecting and extracting highly sensitive meaning from the data with AI is likely to become even simpler and a bigger part of our reality.

With this in mind, I believe it’s important that people’s views and rights are considered. It is the only way decisions about what data can and should never be used, and in what circumstances, can be democratised.

Current laws concerning physiological data collection in the UK and worldwide are inadequate. In the UK privacy laws focus on biometric and neural data, and don’t cover the cognitive and emotional meaning that AI can extract from it.

As I discovered during the exhibition, it is data with mental and emotional meaning that carries the most sensitivity.

‘Bristol in Brain Rhythms’. Viewers put on a Muse EEG headband (bottom left) and watched as images of Bristol, Clifton and the Suspension Bridge were controlled by different brain frequencies.

A solution would be to extend current data privacy principles to cover all cognitive and emotional meaning that AI can extract from data collected from the person. Those principles are (1) informed opt-in consent, (2) data minimization and purpose limitation, (3) rights to access, correct, and delete data, and (4) strong data security including end-to-end encryption. I think legislation is urgent to keep pace with technology and mitigate potential risks to people, e.g. psychological manipulation, harmful disclosures.

The two main ingredients of psychophysiological monitoring — sensitive sensors that collect the raw data, and AI that extracts meaning from it are evolving rapidly. To avoid a global car crash of privacy invasion and personal exploitation by unregulated data-collectors we need to act promptly to put the legal safeguards in place. Otherwise distrust of authorities and corporations with access to this data is justified, and we can’t be sure innovation in personal sensing won’t be to our cost.

“Sensing direct thoughts feels very intrusive and makes me worry about how I could be manipulated when my most private information is easily read.” — Exhibition visitor.

Although my research comprised an ad hoc survey of art gallery visitors, it did indeed show how people were most accepting when they voluntarily provided the information, and least when the information contained personal meaning, especially at the level of thought.

For now, I am grateful to everybody who came to the exhibition and shared their opinions. I’m planning to further explore public attitudes at my next exhibition; venue TBC.