What you will learn
Find out what benefits AI can offer the public sector and how it can help improve the lives of citizens globally.
What you need to do
Read the article (15 minutes)
Complete the reflection activity (5 minutes)
Search across all content
Find out what benefits AI can offer the public sector and how it can help improve the lives of citizens globally.
Read the article (15 minutes)
Complete the reflection activity (5 minutes)
In the last lesson, you learned that when AI is used to make predictions about individuals or groups of people, it comes with a greater risk of causing harm.
In this lesson, we’re going to restore AI’s reputation and get you feeling good about its potential to improve the lives of citizens and the public service.
Let's revisit the list we introduced you to in the last lesson. (We've removed the high-risk examples that involve making predictions about people.)
Take a moment to consider the remaining examples: do they have anything in common?
Manage procurement processes
Optimise refuse collection
Review air cargo records for threats
Predict energy consumption
Respond to local residential issues
Monitor the quality of medical supplies
Transcribe parliamentary meetings
Manage traffic flow
You know, now, that AI interventions are less suited to scenarios that depend on contextual information requiring human observation and judgement. On the other hand, some areas of the public sector are well-suited to AI applications.
In scenarios where understanding contextual information isn’t essential — like in ‘physical’ systems where reliable patterns exist, such as managing traffic flows or predicting energy demand — AI is particularly good at forecasting quantities, labelling categories, predicting likely outcomes and identifying patterns.1
But this doesn’t mean the remaining potential uses of AI are risk-free; you’ll learn more about weighing up risks and benefits in the next lesson.
When writing this course, we discovered that positive examples of AI being used in the public sector far outnumbered the negative examples we were prepared to find. Honestly, we’re as surprised as you are, but there are some things to keep in mind.
Before we accelerate to chase AI’s potential, we’re going to pump the brakes, exercise due diligence and recap Case 1 from the previous lesson. As you remind yourself of the potential for bad outcomes, consider the timescales involved:
2004: The Dutch tax authorities start using AI to detect misuse of its childcare benefits scheme.
2005—2019: Childcare allowance is illegally stopped for thousands of low- and middle-income families. Families are also forced to pay back large sums of money.
2020: The Dutch government begin compensating the impacted families.
Key learning: It took the Dutch government 16 years to start rectifying the damage caused by implementing an AI 'solution'.
Positive examples are easy to find and this helps place fears around AI into perspective. But, negative impacts can culminate over long periods of time.
The achievement of these innovations hinges on AI technologies being built responsibly — slowly, safely, securely and in the public interest.2 – Tabitha Goldstaub, Former Chair of the UK government's AI Council and author of ‘How to Talk to Robots’.
AI is developing rapidly, and it’s important that governments are swift to learn and respond to these technological advancements. Keeping pace with AI developments is a good thing. But this shouldn’t be confused with implementing AI technologies at speed, just for speed’s sake. Caution and forethought need to be the top priority.
Have we succeeded in slowing down your enthusiasm for using AI in government? Well, buckle up! We’re not going to let this pause for thought grind us to a halt. We promised you’d discover a world of potential benefits and we’re ready to deliver.
Below are two case studies that demonstrate the power of AI to help solve problems that result in good outcomes and tick the ‘slow, safe, secure and in the public interest’ boxes.
When you’re ready to be inspired, navigate to the expandable boxes below.
Use case: Review air cargo records for threats
Country: Canada
Years: 2017—ongoing
Each year, Transport Canada receives nearly one million pre-load air cargo records. Each record contains up to 100 fields of information, such as sender and recipient names and addresses, the cargo weight and the total number of packages in each shipment. One employee, working at an unrealistic rate of one record per minute, would only have time to review a fraction of these.3
So in 2017, Transport Canada began a pilot project to see if AI could automate the risk assessment process. The team used machine learning to analyse historical data and identify patterns that could indicate risk (the ‘bomb in a box’ scenario), also developing a natural language processing (NLP) tool to automatically tag cargo records with risk indicators.4
The results of the pilot project were successful. AI could identify potential high-risk shipments with a high degree of accuracy and did so much faster than manual assessments. It is estimated that, “because every single cargo record can be addressed, instead of the small subset possible with manual assessments, AI has the potential to increase safety and security 15-fold.”5
Transport Canada is taking steps to finalise their innovative system.6
Use case: Monitor the quality of medical supplies
Country: Germany
Years: 2019–ongoing
In 2017, the World Health Organisation estimated that 10.5% of the medications worldwide are subpar or fake, substantially impacting the health and wealth of peoples and nations globally. 7
The pharmaceutical industry, regulators, public health and customs have checkpoints for assessing medical products. However, due to the complexity of the supply chain, counterfeits still find a way into distribution. Consumers, re-sellers and hospitals have no instant and global means of checking the quality of medical goods.8
TrueMed is a private-sector-led pilot project that uses computer vision to detect counterfeit medications via a smartphone app. Currently being tested for its accuracy and capacity, TrueMed envisages a successful roll-out of its product as a ‘win-win’ for governments and citizens.9
This ideal scenario is far from being a reality, and there would be many risks and challenges to address along the way, but let’s consider the enormous potential presented by an innovation like this one. As well as helping to detect counterfeit medication at a domestic level, the data collected and interpreted could help international governments identify where counterfeit products were coming from. Successfully using AI in this case would lead to protecting the health of the public worldwide.10
Your time is precious, so we’ve picked just two examples to get you thinking. If you’re interested in reading more about the specific case studies we’ve used in this lesson, check out the Sources list at the bottom of the page.
You can also discover more exciting examples by heading straight to The Observatory of Public Sector Innovation (OPSI) case study library. The OPSI collects and shares hundreds of government innovations to help disseminate and replicate good ideas. Currently, the platform hosts more than 800 cases in total, over 100 of these cases involve AI.11
As the case studies in this lesson demonstrate, using AI in the public sector presents opportunities for solving problems that vastly benefit citizens. But — as you learned in the previous lesson — care and reasonable caution is needed to prevent harm.
Head to the next lesson, where you’ll learn more about navigating AI risks and rewards.
Areas of the public sector where reliable patterns exist, like predicting traffic flows or energy demand, are well-suited to AI applications.
There are many cases of AI being tested and developed responsibly in the public sector.
High-risk use cases can be mitigated when transparency and human oversight such as risk assessments are built into AI solutions.
AI offers vast opportunities to help improve the lives of citizens globally.
Waller, Paul. “Algorithms and AI in the Public Sector: The Rules.” Apolitical, June 17, 2022. https://apolitical.co/solution-articles/en/algorithms-and-ai-in-the-public-sector-the-rules.
Luckin, Rose, Karine George, and Mutlu Cukurova. AI for School Teachers. CRC Press, 2022, p. vii.
Berryhill, Jamie, Kévin Kok Heang, Rob Clogher, and Keegan McBride. “Hello, World: Artificial Intelligence and Its Uses in the Public Sector.” OECD Working Papers on Public Governance. Organisation for Economic Co-Operation and Development (OECD), November 21, 2019. http://dx.doi.org/10.1787/726fd39d-en, p. 149.
Observatory of Public Sector Innovation. “Artificial Intelligence and the ‘Bomb-in-a-Box’ Scenario: Risk-Based Oversight by Disruptive Technology,” October 25, 2018. https://oecd-opsi.org/innovations/artificial-intelligence-and-the-bomb-in-a-box-scenario-risk-based-oversight-by-disruptive-technology/.
Berryhill, Jamie, Kévin Kok Heang, Rob Clogher, and Keegan McBride. “Hello, World: Artificial Intelligence and Its Uses in the Public Sector.” OECD Working Papers on Public Governance. Organisation for Economic Co-Operation and Development (OECD), November 21, 2019. http://dx.doi.org/10.1787/726fd39d-en, p. 149.
“Canada Gazette, Part 1, Volume 157, Number 11: Regulations Amending the Canadian Aviation Security Regulations, 2012 (Air Cargo).” Accessed August 18, 2023. https://www.canadagazette.gc.ca/rp-pr/p1/2023/2023-03-18/html/reg5-eng.html.
Pathak, Ranjana, Vaibhav Gaur, Himanshu Sankrityayan, and Jaideep Gogtay. “Tackling Counterfeit Drugs: The Challenges and Possibilities.” Pharmaceutical Medicine 37, no. 4 (July 2023): 281–90. https://doi.org/10.1007/s40290-023-00468-w.
Observatory of Public Sector Innovation. “Rapid, Optical and Global Detection Tool against Counterfeits in Every Pocket,” October 20, 2021. https://oecd-opsi.org/innovations/rapid-optical-and-global-detection-tool-against-counterfeits-in-every-pocket/.
Observatory of Public Sector Innovation. “Rapid, Optical and Global Detection Tool against Counterfeits in Every Pocket,” October 20, 2021. https://oecd-opsi.org/innovations/rapid-optical-and-global-detection-tool-against-counterfeits-in-every-pocket/.
Observatory of Public Sector Innovation. “Rapid, Optical and Global Detection Tool against Counterfeits in Every Pocket,” October 20, 2021. https://oecd-opsi.org/innovations/rapid-optical-and-global-detection-tool-against-counterfeits-in-every-pocket/.
Observatory of Public Sector Innovation. “OPSI Innovations.” Accessed August 21, 2023. https://oecd-opsi.org/case_type/opsi/.
⬅️ Previous lesson | Course outline | Next lesson ➡️
This is the logged out version of a short course in Apolitical’s Government AI Campus. To access all interactive elements and download a certificate of completion, sign up for free.