“Smart cities” pose fresh ethical challenges for open governance
By: Alex Howard.
Where do “smart cities” meet open cities? Technology is implicated in both, but the forces in society that back each of the visions are often radically different. The role technology can, should, or will play in cities is a tension that ran through the Open Cities Summit on Monday, as advocates grappled with ongoing wave of automation and sensorization in municipalities around the globe. Technology companies investing in smart cities aren’t always embracing openness by default in their operations or the data that they will produce.
Eric Reese, associate director of Johns Hopkins’ Center for Government Excellence, suggested to the Open Cities Summit that the key thing for city officials and public servants to do is to manage relationships among people in a way that optimizes for trust.
To him, a “smart city” is full of smart people who have access to information and can use it to improve lives, in which a local government works collaboratively to solve problems and empowers the public with the skills they need.
The challenge is that that the “smart city” agenda has been driven primarily by huge corporations and involves adding sensors, surveillance, and analytics throughout cities. While improved services, efficiency, and even lower mortality are potential outcomes, so are huge risks to human rights and civil liberties — and that these smart systems are often being procured and installed without the consent of the governed.
As more services become digitized or are born digital, the ethics embedded in the code has also become a paramount concern, as Milena Pribic, a researcher IBM explored at the summit. Algorithmic transparency, accountability and ethics are no longer theoretical constructs but core to democratic outcomes in a networked world.
Her work at IBM, which where she studies the relationship of humans with artificial intelligence (AI), focuses on how ethics relate to public participation and transparency.
Pribic suggested that cities, states and nations — and the companies that build services for them — need to invest in “democratized AI”, which she defined as a participatory system that operates on majority rule and individual rights, paralleling the structure of the city itself.
The ethical foundation for AI that she outlined has five pillars:
- Accountability, where everyone involved is accountable for considering the impact of the system on the world.
- Value alignment, where algorithms reflect the values of a given community and can be changed or adjusted if those values change
- Explainability, where services are designed so that humans are able to perceive, detect, understand a decision process. She recommends a “nutrition label” for AI, disclosing category, purpose, reasoning, risks, and potential bias.
- Data rights, which protect data & preserve a person’s power, avoiding stripping someonsmae’s identity of context
- Fairness, where services are designed to minimize bias and maximize inclusion. She notes that this starts on the team that builds a service, contrasting diversity with monoculture and noting that equal representation is not inclusivity — need equal participation
“There’s a market advantage in building trust”, she said, arguing that when we’re thinking ethically, thinking inclusivity, there’s bias reduction, scalability of systems and services, and scalability of relationships.
Mayors, advocates, technologists and all of the other stakeholders in communities will all need to think through how and were democratic governance intersects with modern technologies, including those that appear benign and are implemented with the best of intentions. Doing otherwise could jeopardize the work and legitimacy of open government advocates, all of which are in play in societies around the world.
That’s a risk that merits mitigating. You need the support of the people, who need to understand government is not something foreign to them, said Tony Rojas, a data scientist at Puebla Rojas, responding to a question about public involvement and engagement.
It is necessary for us to become active, not passive, citizens, Rojas said.
“If we don’t create a network of data generators and participators, not consumers, it’s difficult to see how smart cities flourish”.
