Open and Transparent Government

 Download the report (PDF)

This document does not represent an official policy position of the Government of Canada. Instead, it records the work of a sub-group of new public servants who participated in Canada Beyond 150, a professional development program co-championed by the Privy Council Office and Policy Horizons Canada. The program was designed to support the development of new public servants, and to drive a culture change within the public service. The participants were invited to use foresight, design thinking and engagement tools to explore policy issues relating to diversity and inclusion.

Introduction

When you hear the terms “open” and “transparent” government, what comes to mind?

Do you think of a government that is open to the people it represents and serves? Or about transparent decision-making and accountable decision makers? Or do you think of a government that connects citizens and helps them participate in government processes?

To prepare for the foresight analysis, the team reviewed Canadian, foreign and international background documents relating to open and transparent government. The team also identified a number of weak signals—or emergent developments that might predict larger disruptive changes across government. Some weak signals are included under each of the policy challenges outlined below.

As part of the foresight and policy intervention design processes, the team interviewed about 50 stakeholders and partners from a wide range of sectors: federal government departments, provincial and municipal governments, academia, media, private sector companies, innovation think tanks, and non-government organizations. With expertise in areas such as service delivery, technological design, data sharing and privacy, and citizen engagement, these interviews added rich perspectives to our work and helped develop our understanding of the landscape of open and transparent government and its potential evolution.

Policy Challenges

Algorithmic Bias and Transparency

"The next frontier for freedom of information laws around the world is code. Whether code is used by a city, state or nation, it’s incumbent on public officials to avoid prejudice in data-based decisions." – Alex Howard, Sunlight Foundation

Algorithms have the potential to dramatically improve the provision of services, administrative processes, sharing of data, and the evidence base for decision-making. However, such processes can reflect human biases and present accountability challenges. Given the scale of government operations, algorithms could result in systems that implement biased or opaque decisions on a previously unimaginable scale and could undermine a citizen’s right to procedural fairness.

For example, in the US, a risk assessment tool that was used to rate the likelihood of defendants to commit future crimes was found to show racial bias. ProPublica analyzed sentencing and found that black defendants were incorrectly labelled as future criminals at almost twice the rate of whites. This problem raises concerns over the use of other secretive, proprietary ‘black box’ algorithms in determining such fundamental decisions for citizens.

In another example, issues around diversity in algorithm design were highlighted when facial recognition software could not identify people with darker skin tones. In this case, the exclusion of minority groups from algorithm training data sets meant that the resulting product and service was not able to serve all its users equally. This reinforces questions of who benefits from new technology and how.

When considering the government landscape over the next 15 years, the potential benefits of using algorithms make it highly likely that government will rely on algorithms to deliver services and make decisions. However, the success or failure of this increased use will depend on how the government responds.

If algorithms become ubiquitous, it will become fundamental to ensure that data and code are transparent and unbiased to Canadian democratic institutions. Algorithms present a global challenge that will affect other governments and the private sector; Canada has an opportunity to become a global leader in the responsible use of algorithms.

The proposed interventions for this policy challenge are driven by the following assumptions:

• Algorithms will be implemented to automate and support administrative decisions (short term) and support analysis and research to develop policy (long term).
• Both humans and algorithms have the potential to be biased and to lack transparency. However, compared to people, a single algorithm has a greater potential to implement problematic or biased decisions on a vast scale, so there is a stronger duty to ensure procedural fairness.
• The development and use of algorithms in this setting is very new, and it is too soon to implement any strict policies that are not adaptive to the changing landscape.

Interventions

Immediate term (1–2 years): Laying the foundation for a digital future

• Proactively acknowledge that Canadians’ equality rights require protection against discrimination embedded in algorithmic decision-making.
• Create Canada Research Chairs on Algorithmic Fairness and Algorithmic Auditing, involving both computer science and social science perspectives.
• Through talent management and training, develop the Government’s capacity to develop and deploy algorithms, and to understand their consequences on diverse populations.
• The Government should develop principles-based guidelines on the responsible use of artificial intelligence (AI) and decision making in program management, which the Treasury Board of Canada Secretariat has started working on.

Medium term (3–5 years): Institutional response

• Create a parliamentary position of Algorithmic Ombudsman responsible for enforcing AI transparency.
• Create an algorithmic evaluation service available to governments, and potentially companies, across Canada.

Access to Information

Access to information is at the heart of open and transparent government. The state is only transparent if people can find out what government is doing. However, Canada’s access to information system is often seen as failing to provide meaningful transparency, while imposing excessive costs. Improvements in information technology could start to address these concerns, but they will not be perfect cures.

The problem begins with the ‘juiciest’ documents, such as cabinet materials, often being exempt from disclosure that under the Access to Information Act. When a document may be disclosed, it is often released late and in redacted form. Meanwhile, the system is troublesome for public servants, who have to spend hours compiling and redacting old emails that are ultimately often of little interest to the person who made the request. These problems can only be expected to get worse: the number of access requests continues to rise by 10 percent each year (as of 2016), government has more and more information, backlogs are growing and no one is happy.

Early in the project, the team hoped that the trend towards proactive disclosure—the practice of releasing documents before they are requested—could address these concerns. Some stakeholders in media and civil society suggested that more systematic disclosure might reduce the demands on the access system, while enhancing transparency. However, we heard from experts within government that their experiments with proactive disclosure generally led to more document requests: when a document is published, people request more related documents and correspondence. We also heard that official languages and accessibility requirements make it very costly for departments to publish as much as they would like.

The team experimented with design thinking to try to find another solution. A public registry of government documents along the lines of the Norwegian eInnsyn (e-Access) system could make it easier for people to download or request government documents, and perhaps eventually reduce the administrative burdens of request-based systems. The Open Government portal already lists some documents, including ones that have been previously released in response to requests.

We propose interventions for this challenge, based on the following assumptions:

• Few people access documents, even when they are easily available.
• Many people’s understanding of government is mediated by news and social media, which are often more interested in covering conflicts and error than earnest policy making.
• Improvements to the access system probably will not change how people think of government.

Interventions

To take advantage of these challenges, government should:

• Plan with the future in mind:
  • Government can adopt practices and procedures that lend themselves to future automatization. For example, officials could write documents with access in mind, keeping confidential information in specific sections of reports, to make it easier for software to determine what can be disclosed.
  • Establish quality standards for machine translation in advance, so that the Government can take advantage of new tools as soon as they are available.
• Develop a stronger interface between open government projects and access to information offices, and improve internal information management systems in order to identify the documents of greatest public interest and prioritize them for disclosure.
• Use technology to alleviate administrative burden:
  • Machine translation could allow electronic publication of documents in both official languages. (The previous sentence was translated from French by Google.)
  • Software algorithms could also assist, and eventually take over, the process of compiling and redacting documents.

On the other hand, we need to be cautious about thinking of technology as a cure-all. While better access systems would support the public’s “right to know,” they are unlikely to do much to encourage public trust in government (and broader open government goals).

Digital Divide

Open dialogue, social inclusion and civic participation are at the heart of open and transparent government. But while new divides are emerging around digital literacy and access to services, another digital divide is threatening the social fabric and security of Canada and other Western nations. In this digital divide, groups of people are separated by fissures in social cohesion and breakdowns in public discourse. The challenge, for government in particular, is to determine its responsibility in fostering the free and open dialogue that sustains public trust and social cohesion.

Iconic movements like #IdleNoMore, #BlackLivesMatter, #MeToo and Arab Spring have given a voice to traditionally marginalized communities. Citizens are directly engaging with politicians and decision makers in ways that would be unimaginable even 10 years ago. However, while these digital spaces provide platforms for citizen engagement, economic innovation and mobilization, they also enable the creation of echo chambers where like-minded people can isolate themselves from exposure to competing ideas and perspectives. Over time, this can contribute to a breakdown in social cohesion and public discourse. These fissures in the social fabric can have wide-ranging impacts, including adverse effects on our electoral and government institutions.

We proposed interventions for this challenge, driven by the following assumptions:

• Citizens will increasingly continue to use digital platforms and spaces as platforms find new and innovative ways of attracting and retaining users.
• Increased use of these digital platforms will have implications for real-life interactions, dialogue and social cohesion.
• The Government has a role in supporting free and open expression and communication online.
• The Government also has a role in sustaining social cohesion and public trust.

Interventions

Immediate term (1–2 years):

• Collect baseline data on the state of digital engagement and digital literacy in Canada.
• Create a royal commission to engage with Canadians, businesses and experts in a national dialogue on digital freedoms in a connected age.
• Sponsor an XPRIZE-style competition to encourage cross-fertilization across digital divides, and motivate those with technical expertise to develop innovative solutions to resolve technical issues exacerbating the digital divide.

Medium-term (3–5 years): Institutional response:

• To outline a response to the social harms, develop a public health and social well-being approach to the digital divide with specialists (e.g. epidemiologists; social workers; curriculum developers and educators; journalists; technical experts).
• To help government prepare for emerging digital trends, monitor the emergence of new digital divides in collaboration with academics, researchers and other stakeholders.

Lessons Learned

Project Management in a Complex, Organic Team Environment

The team was composed of 16 members, and initially exerted substantial time and effort to establish project management processes rather than focus on the subject material. As a result, we did not have as much time as we would have liked to spend on some topics. Some topics, we were not able to address at all (see examples below). To define the project scope more clearly, the group used an ‘accordion’ style of work. The goal was to divide into three distinct teams to work on individual theme areas, and then come back together to share information with the entire group. The three themes—disclosure, engagement, and service to vulnerable populations—were selected based on environmental and weak signal scanning.

Engagement with Partners – Not as Open as Planned

During the final months of the project, the group established an outreach team with the goals of identifying gaps in engagement, facilitating sessions with stakeholders, and implementing a strategy to re-engage stakeholders from earlier phases of the project. Remaining in contact with stakeholders was an ongoing challenge. As well, due to security concerns, unequal access to the shared document tool across government departments, and short timelines for approvals, meant the group was forced to adjust plans and to contact stakeholders by email only. Our team tried to engage stakeholders using their preferred methods and platforms (like Google Docs and Slack) to increase the likelihood of response. However, even within a program like Canada Beyond 150, designed to push the boundaries and experiment, the team was not immune to common constraints placed on most government processes.

For example, the team designed an engagement activity to test policy interventions on access to information (ATI) and privacy. The activity consisted of a workshop to invite journalists and ATI officers inside government to create a journey map of their experience using the ATI system. Officials guiding the project advised that the team needed to first consult communications and ATI specialists at the lead department. Given the timeframe was quickly shrinking, the team decided to cancel the original plan as it would not have enough time to consult the necessary parties beforehand.

Reaching the Hard-to-Reach

The Outreach team also sought to include the views of “hard to reach” stakeholders, such as newcomers, youth and individuals from low socio-economic backgrounds. In doing so, the team learned that engagement with these stakeholders is a long-term process that requires building trust over time. While the team made efforts to reach out to four different organizations representing newcomers, youth and other marginalized populations, only one youth-serving organization was able to support a discussion on open and transparent government on the margins of their conference. Unfortunately, this engagement came too late in the project to adequately reflect on how these youth perspectives could be included in our analysis. Similarly, a city councillor with a riding in a low socio-economic neighbourhood showed interest in speaking with the Outreach team, but again, late timing did not allow us to schedule an interview for this project.

Conclusion

Our team’s journey exploring open and transparent government was rich and rewarding. We feel the policy challenges we developed are worthy of further development and hope that we, with others, will have opportunities to pursue them.

Team

Enabler

Canada Beyond 150 was made possible through the support of hundreds of stakeholders and partners whom were engaged over the 10 month journey.

Top of the page