Embracing technology: the public sector of the future

To better serve its citizens, the public sector faces an existential need to become more agile, more mobile and more efficient. Some of the most hotly anticipated solutions include those enabled by artificial intelligence (AI). Ranging from predictive analytics to machine learning to intelligent robotic process automation, AI is one of the surest paths for extracting insights and value from growing volumes of data.

This has fuelled aspirations for everything from advanced smart cities to new approaches in population health management – often these solutions involve predictive analysis that could help agencies make better decisions, respond faster during crises and even pre-empt problems altogether. Some agencies are making use of AI applications already, likeĚý, which used machine learning to predict tax non-compliance and netted the state an extra $27 million in revenue.

Government also has a unique role to play when it comes to AI – since all Australians are impacted in some form or other by government services, governments must take the lead in their use of AI, whether through operations or service delivery.

Yet broader adoption remains low. A 2018 investigation by the 51·çÁ÷Institute for Digital Government (The SIDG) found that, while 80 per cent of public sector organisations were working toward data transformation, less than 15 per cent had progressed beyond the prototype stage.

The SIDG teamed up with University of Queensland researchers to assess where the sector is at in 2020. The resulting white paper,Ěý, identifies the biggest AI challenges in the public sector – and how leaders can overcome them to finally harness the true potential of AI solutions.

The resource challenge: building AI capability and securing human talent

AI relies on large datasets, high-quality data, the right platforms and – importantly – data science talent.

This is resource-intensive – an acute challenge in the public sector where data is often purposefully siloed, and fractured across complex, ageing legacy systems. These overlapping issues create a sort of chicken-egg dilemma, where leaders may struggle to secure funding and executive buy-in without proven value – but proving value depends on funding and executive buy-in.

The research did uncover examples of success, though. One agency was able to overcome data-sharing barriers by outsourcing its AI model development, which was then trained with citizens’ payment data instead of sensitive personal data. Another agency chose a commercial-off-the-shelf AI development platform to decrease maintenance burdens.

Misunderstandings about AI and inflated hopes also demand project-level governance to manage expectations and encourage ongoing commitment from executives.

The process challenge: pre-empting machine fallacies by keeping humans in the loop

Despite myths of robot overlords and job losses, algorithms only outperform humans in their ability to process huge datasets. They still lack the context-specific reasoning capabilities that we have, which means AI solutions can’t simply be plugged into existing workflows. Agencies will need to rethink processes to combine the strengths of machines and people.

This is complicated because of the barriers that often separate data scientists and subject matter experts, demanding redesign for entire workflows. The researchers found that agencies who were able to reconcile these issues were those who embedded data scientists in everyday operations and encouraged collaboration with subject matter experts.

Successful approaches include co-location and collaborative workshops but, interestingly, interview data also highlighted the importance of attracting data scientists with strong soft skills and good communication.

Organisations were keenly aware of the need for human oversight and the risks of deferring to automation. Many were already redesigning workflows to ensure AI was doing the heavy lifting and data-crunching, with human workers acting as the controllers of the AI and making final decisions.

The explainability challenge: minimising bias and enabling transparency

Advanced AI models have an “explainability problem” – that is, the complexity of their logic and the sheer volume of data can make decision-making inscrutable to us.

This is a massive hurdle in the public sector, where public trust often depends on transparent rationale and straightforward accountability. It’s an even bigger challenge once we consider that algorithms have already demonstrated a serious risk of bias and error.

The researchers found that some agencies have been establishing strict oversight and procedural systems with these specific risks in mind. For instance, one agency excluded demographical data in favour of behavioural data to minimise bias in the model’s predictions.

Another created a more extensive end-user interface that visualised a customer journey and highlighted risky payment behaviours. This provided visibility into the factors affecting the overall risk estimate.

The culture challenge: reducing distrust among employees and citizens

Despite research indicating AI adoption rarely comes from a desire to reduce headcounts, job security fears abound. Additionally, the researchers found some human workers continuing to distrust AI’s decisions.

One solution is educating employees about the potential of AI-enabled tools – this can be an easier sell once employees witness the elimination of low-value tasks and admin burdens, freeing them to focus on more strategic and interesting work.

The public sector faces public resistance, too. Some agencies have the added challenge of a power imbalance, as citizens who rely on their services may not be able to switch providers like they would in the private sector.

While wider societal perceptions may evolve in a way that reduces distrust, there’s no simple solution to these challenges. Trust will depend on proven value and the effective management of unintended consequences – which will in turn depend on many of the solutions mentioned above.

The public sector faces unique challenges with AI solutions but also stands to gain some of the biggest rewards. And, promisingly, some agencies are already demonstrating how to address these issues.

Using an even deeper look into the public sector’s relationship with AI,ĚýĚýprovides a practical framework for developing the foundations necessary for effective AI development in government.

However, it’s an area that requires deeper exploration, which is why The SIDG will continue partnering with the University of Queensland to understand ongoing challenges.

To read more about 51·çÁ÷Australia’s public sector offer, Ěý

The post Government has yet to fully capitalise on AI. Here are 4 ways to change that. appeared first on 51·çÁ÷Australia & New Zealand News Center.

The first two articles in our “Dealing with Disruption” series looked at how digital technologies might enable governments around the world to nudge citizens towards cooperation and coordinated action in containing COVID-19, and to address issues of hand washing, face touching, self-isolation, collective action, and crisis communication. In this article, the 51·çÁ÷Institute for Digital Government (SIDG) will present a conceptual architecture for Digital Nudges and demonstrate how it could be applied to improve crisis communications relating to a second-wave outbreak of the Coronavirus.

Using digital nudges to support government responses to coronavirus

To demonstrate how our conceptual architecture might be applied, we will consider the scenario of a second-wave outbreak of the Coronavirus, such as was .

Figure 1: The first- and second-wave outbreaks of COVID-19 in Australia.

was identified on 25 January 2020. The number of new cases rapidly increased and peaked nine weeks later, with reported on 28 March. The Australian government responded very successfully with a for flattening the curve, and by mid-April there were a relatively low number of new cases being reported daily. Although the virus had not been eliminated, it appeared to have been suppressed sufficiently for lockdown restrictions to be eased across Australia. Unfortunately, were identified in Melbourne on 20 June, foreshadowing a second-wave and prompting a reinforcement of restrictions to contain the outbreak. Even so, Australia’s second-wave proved more difficult to contain than the first, peaking at reported on 5 August.

Due to the localized nature of the second-wave outbreak, stay-at-home restrictions were reintroduced only in metropolitan . Most notably, in North Melbourne and Flemington were immediately locked-down, with residents of 33 Alfred Street subsequently required to isolate for two weeks. While it was generally agreed that this was a necessary measure, the immediacy of the action combined with various communication challenges resulted in widespread confusion and concern among the 3,000 public housing tenants. captured the sentiment at the time:

• “When I came back home I did see hundreds of cops everywhere, so it was really intimidating.”
• “It’s been getting more and more intense, people are really panicking.”
• “We weren’t told any information, they just shut us down, didn’t let us leave our houses.”
• “I just feel like we’re being treated like criminals.”
• “We do not need 500 officers guarding the nine towers. We need nurses, we need counsellors, we need interpreters.”

In what has been an unprecedented year, the hard lockdown of Melbourne’s public housing towers was an unprecedented action by the Australian government, law enforcement and public health services. To that point, Australian citizens had not experienced a lockdown under guard, except in cases of returned citizens undertaking hotel quarantine.

In special cases such as this, efficient and effective crisis communication is key – not only in ensuring compliance – but in promoting cooperation through credibility, empathy and respect. Behavioral Science can assist by influencing individual decisions towards the most positive outcome, and digital technologies can be used to scale and personalize traditional nudges to improve outcomes for mass cohorts.

Conceptual Architecture for digital nudgesĚý

Figure 2: A conceptual architecture for digital nudges.

Nudging is a delicate process, with significant preparation required to avoid unintended consequences – especially when the stakes are as high as they are in the case of COVID-19. These stakes are raised even higher when the nudges are to be delivered by governments, at scale, using digital technologies. The is to optimize utility and mitigate risk using an iterative process of randomized controlled trials with rapid cycle evaluation. Whether the nudge is to be delivered as part of a trial, or to the population at large, an iteration of the nudging process typically spans:

• Design and contextualize: The nudge is designed to achieve the outcome of interest, based on an exploration of the available data. A key consideration is the situational and social context of the environment in which the nudge is to be deployed. In the case of crisis communications, nudges need to for citizens’ circumstances.
• Simulate and deploy: Randomized controlled trials can be used to simulate the likely response to a given nudge. A variation of this approach would involve using , to enable simulations to be run faster and safer than with human subjects. In the case of crisis communications, these simulations could be aligned to the accepted thresholds of a national or local containment strategy.
• Monitor and measure: Having deployed the nudge, social listening and devices can be employed to monitor the actual response. Although it may be difficult to measure the effectiveness of nudges as a behavioral modifier, a control group who does not receive the nudge may be used. In the case of crisis communications, we might also consider performance against “fake news” as a measure of effectiveness.
• Analyze and improve: Here we distinguish between measurement and analysis, specifically within the context of diagnostics – analyzing why a particular action has been taken or a particular outcome achieved. Based on this analysis, improvements can be made to the design of the nudge, and thus the iteration continues. In the case of crisis communications, certain visualizations (e.g. ) might be published to encourage community cooperation and coordinated action.

Digital nudges: Core capabilities

As described in our first article, predictive analytics, contextualization, and experience management are the core capabilities required to deliver digital nudges. Breaking down these capabilities will enable us to illustrate how they can support policymakers and service agencies, working with behavioral scientists and technology partners, to improve the effectiveness of traditional nudges.

• Predictive Analytics:
  • Behavioral Insights: The ability to detect patterns in citizen behavior, based on transactional and experiential data. For example, based on their prior responses to government requests, we can expect Citizen X to comply with stay-at-home orders.
  • Journey Visualization: The ability to visualize the citizen’s journey over time, including major life events, changes in circumstance, and their interactions with government. For example, based on the healthcare, social services and financial supports they have recently accessed, Citizen X is likely a vulnerable person who will need additional supports.
  • Simulation: The ability to simulate the likely responses to a digital nudge, including the ability to compare alternative approaches. For example, Nudge A will increase compliance with stay-at-home orders by 5%, with 80% confidence.
  • Next Best Action: The ability to recommend the optimal course of action, based on (autonomous) machine learning. For example, Nudge A will be most effective for Citizen X, while Nudge B will be most effective for Citizen Y.

• Contextualization:
  • Profiling: The ability to assemble a digital profile of a citizen, by combining data from multiple sources (as permitted by government regulations). For example, we know that Citizen X is at high risk, since they are over 80 years of age and live in high-density public housing.
  • Segmentation: The ability to create target groups, comprising citizens with similar profiles and needs. For example, Segment A comprises citizens of working age, who are likely concerned about the impact of stay-at-home orders on jobs.
  • Campaigns: The ability to proactively outreach to target groups with nudges tailored to their circumstances. For example, Nudge A will be delivered to citizens of working age, while Nudge B will be delivered to citizens over the age of 65.
  • Preferences: The ability to communicate with citizens via their preferred channel, and at their preferred time and place. For example, Citizen X usually responds promptly to SMS sent around lunchtime.

• Experience Management:
  • Social Listening: The ability to monitor social media to track changes in citizen sentiment over time. For example, citizens under lockdown are complaining that police presence is making them feel like criminals.
  • Surveys: The ability to solicit direct feedback from citizens. For example, Citizen X responded that they couldn’t understand the specifics of the stay-at-home order because English is their second language and no translation service was provided.
  • Measurement: The ability to measure the response to a digital nudge, based on transactional and experiential data. For example, Nudge A increased compliance with stay-at-home orders by 3%, compared with the control group who did not receive the nudge.
  • Diagnostic Analytics: The ability to uncover why certain nudges are, or aren’t, working. For example, Nudge A was widely criticized as being disrespectful, resulting in a lower level of compliance than anticipated.

The underlying business platform supports the design, development, and management of our digital nudges.

• Analytics: The ability to analyze transactional and experiential data. Desirable features include the ability to:
  • surface actionable insights based on predictions;
  • dynamically drill-down into records of interest;
  • visualize citizen journeys over time; and
  • update data and visualizations in real-time.

• Intelligent Technologies: The ability to build, execute and manage machine learning applications. Desirable features include the ability to:
  • process big data holdings to build advanced machine learning models;
  • support profiling and segmentation of data in line with contextualization capabilities;
  • generate predictions and next best action recommendations; and
  • make improvements based on (autonomous) machine learning.

• Data Management: The ability to access and work with big data, in real-time. Desirable features include the ability to:
  • consolidate data from multiple sources;
  • work with transactional data in real-time, without impacting operational systems;
  • work with analytical data in-place, without the need for replication; and
  • ensure the security and privacy of citizen data.

• Application Development & Integration: The ability to develop and integrate business applications. Desirable features include the ability to:
  • accelerate the design and development of advanced machine learning applications;
  • run simulations in support of what-if analysis;
  • support an open ecosystem of development partners; and
  • integrate with external systems (e.g. geographic information systems).

In presenting this conceptual architecture, our intent has been to provide a framework that governments can use to deliver digital nudges. We believe this framework to be general-purpose, while acknowledging that certain scenarios will require additional capabilities. Our chosen use case of crisis communications serves as an illustrative example. Please note that, since this conceptual architecture is vendor-agnostic, the described capabilities could be sourced from any technology provider.

To read more about how digital technology can be used to improve public sector services, visit .

The post Dealing with Disruption: Conceptual Architecture appeared first on 51·çÁ÷Australia & New Zealand News Center.

I spoke with , Public Sector SME at SAP, and , Executive General Manager of Public Services for 51·çÁ÷ANZ, to look at how governments can be more responsive in helping people and businesses recover from COVID-19.

Chris noted the rapid coordination of industry responses throughout the start of this year. “A few of my observations revolve around the speed with which we’ve been able to do things, whether it be businesses or, more importantly, governments,” he said. “A good example is the bushfire recovery agencies breaking down silos and forming a national body to look and manage the bushfire recovery. In relation to that would be the national cabinet – really cutting across those borders and providing a unified and confident front to COVID-19.”

Ryan offered his personal experiences during 2019-2020 bushfire season, which threated the safety of his home and drove his family to seek safety after internet and power were cut. Thankfully, the fire front passed and left his home untouched, but the experience has highlighted the value of real-time data.

“A key area of focus for me is data-driven government, but that experience drove home for me the importance of real-time information to make life-critical decisions,” he explained. “And again, during the COVID-19 crisis, we’re relying on analytics to trace the spread of coronavirus in real time to inform these types of critical policy decisions. We really just need to look for opportunities to embed analytics throughout our business processes and operational systems – make decisions in the moment when it matters most.

“People in crisis really expect all levels of government to work together to provide the leadership and support pretty desperately needed at during these circumstances. I think it’s that realisation that resulted in the creation of the national cabinet, which has really introduced a new level of collaboration in our federal and state governments.”

As our nation faces recession-level unemployment numbers, helping Australians get back to work in response to the Prime Minister’s request for new policy ideas. Chris explained the six themes and recommendations outlined in the paper.

1. Digitalising small and medium businesses, which employ almost half of Australia’s population and are critical to our national economic and social recovery.
2. Building cross collaboration throughout public service and providing the data and technology necessary for creating a most holistic understanding of business lifecycles.
3. Improving efficiency across all business operations within government and creating a more digitally skilled and connected public service.
4. Focusing on investment going into infrastructure projects, particularly at state level, and applying digital engineering.
5. Utilising smart borders, which was important for keeping skilled people moving across borders, and now paramount in keeping people from moving across state borders during this pandemic.
6. Using intelligent technologies and data analytics to drive responses to unemployment.

Chris expanded on the last point, “We need to utilise technology better, bring big subsets of data together that government already owns, but applying and combining that with some new insights on sentiments of bush fire victims – and now Jobseekers – to help create jobs, stimulate the right employment opportunities in the right places and times.

“First, we need to improve data access. Second is data quality, ensuring high quality and consistency. Last is fighting through data fog since the sheer amount of data makes it difficult. Designing and deploying ongoing data collection activity for things like Jobseeker and the businesses engaged in the Jobseeker program will assist Australians getting back to work.”

Ryan highlighted the value of combining operational and experience data, particularly as governments manage the social and economic recovery of different industries and regions. 51·çÁ÷has also prepared a responsive government playbook () that is full of scenarios that agencies can use to be more proactive in understanding what’s happening citizens and why.

“Our research tells us that citizen transaction and process transparency were the leading indicators of trust in government,” Ryan said. “Based on that, we set about creating a responsive government playbook that could be used by agencies, firstly for the drivers into customer satisfaction, and then to formulate personalised responses based on what they’ve learned.”

My guests agreed that governments need to work more collaboratively, across departments and with its citizens to be more responsive and communicative during this crisis, particularly as we face the challenges of unemployment, citizen safety, and industry recovery.

To learn more listen to the complete episode of the Best Run Podcast .

The post Helping Australians Return to Work Safely and Intelligently appeared first on 51·çÁ÷Australia & New Zealand News Center.