Currently reading:
Patrick Sisson (17 January 2018) 'Your city is watching you: How machine learning and "computer vision" will transform our cities", Curbed.
Sisson suggests that machine vision (which has seen half a billion dollars invested in the technology in 2016) and machine learning (which receives between 4.8 billion and 7.2 billion dollars in global annual spending) offer a further leap in the use of surveillance to deepen our understanding of urban data and to decipher patterns to aid in urban planning. Rohit Aggarwala (Sidewalk Labs) posits that '[t]he ability to transmit images into data, without human intervention, is the single most powerful thing'.
Surveillance in the sensorised building
In sensorised buildings, the application of surveillance technology now permits the 'reading' of occupants' emotion. This has particular relevance in new brick-and-mortar retail experiences (given the current limitations of an app - but watch the 'virtual shop' space). Amazon and the Chinese JD.com are providing unmanned stores powered by surveillance equipment and lines of computer code exploring this new area of entrepreneurial surveillance. And it has implications for manned stores: Walmart's Project Kepler utilises a facial recognition system that 'track[s] consumer mood' and determines how the store's human resources should be allocated to to provide assistance to identified customers. Machine surveillance provided by Standard Cognition can undertake a number of tasks normally given to in-house staff, taking inventory and observing shoppers in real time (and from more perspectives than a single staff member is capable of). In the latter, Standard Cognition's AI 'reads' (from across the room) the intentions and activities of a shopper, assessing what the shopper picks up and interacts with, and is able to recognise when the item is actually purchased.
As the technology becomes increasingly more affordable to implement, the commercial imperative to apply such technology will increase, and in turn produced even greater variety of offerings that will become even more 'invisible' to the shopper. Sisson quotes Daniel Davis from WeWork (coworking entity), the aim of the game is to produce a surveilled environment that 'doesn't look like that kind of high-tech space, it looks homely, inviting, welcoming... [with nothing suggesting] a sophisticated process behind it. The space is performing in a way that suits your needs, without overtly being a space of the digital'.
This wariness of these tech designers/ proponents to human hesitation to increasingly sensorised buildings intrigues me, particularly Ann Sussman's thinking on the 'age of biology' in design - where she has used eye-tracking software 'to look at how humans react to architecture and urban design'.
Urban planning level
This emphasis on creating technology that doesn't read as technology, and that still remains attractive to users (benchmark: if occupants lower their mobiles), is also present on the urban planning / smart city planning level. Aggarwala (Sidewalk Labs) notes that 'If it feels tech-forward, we've probably done a few things wrong.'
Surveillance and urban data gathered could be used to help authorities manage a suburb's facilities to meet the needs of its changing demographics (eg changing park facilities to provide for more mature families as the suburb ages etc), or to even implement urban features that increase safety, encourage or attracts particular demographics to a locale.
Challenges
Surveillance technology is limited to how it is trained. Critics note that where such training is imperfect, the likely users of the technology, such as 'women, young children, and African Americans' are likely to affected. This has particular consequences when used for law enforcement purposes (see 'The Perpetual Line-Up', Georgetown Law Center on Privacy & Technology). Sussin notes that there are commercial applications of such technologies already: Ekin, a company that sells patrol cars with facial recognition technology, provides products that will identify (through facial recognition and analytics) 'suspicious, guilty, or wanted people even the human eye can miss...'.
A lighter-touch version of surveillance ('soft surveillance' - where data is de-identified and processed to only produce object classes) - may be preferred where there are concerns about cost and privacy. For example, Numina creates a counter that attaches to street poles that counts bicycle and pedestrians to provide hard data for urban planners when considering the breakdown of road users.
