Pandemics & politics: Can data-driven AI make better policy?

Published by IAPA

Dr Nuria Oliver’s algorithms model the perfect politician, one who can make accurate and cost-effective policy responses amidst the chaos of a COVID-19 outbreak. Could this start a new global trend: evidence-based policymaking?

With the public weary of lockdowns, hospitals overloaded, vaccine shortages and policy-makers drowning in the chaos of a COVID pandemic, Spanish data scientist Dr Nuria Oliver knew that mathematics and big data could make better decisions than a human politician.

Her team got to work building big data models to forecast the incidence of COVID-19 and cost-effective policy interventions to minimise death, economic havoc and spread of the dreaded disease.

The model took 12 different ‘dimensions’ - or interventions such as closing workplaces and schools, mask-wearing and education - so the AI modelling could spit out the very best policy intervention given the hospital loads, economic costs and testing capabilities.

“When you look at the 12 dimensions, there were different activation levels which led to 7.8 million possible combinations,” explains Dr Oliver, who has published hundreds of research papers and developed more than 40 patents for data-driven solutions.

“No politician could likely comprehend those 7.8 million different combinations but a machine can - a machine can find the policy that satisfies the outcome against the given cost so it’s always offering up the optimal policy. It’s evidence-based policy making.”

AI helping humans

Dr Nuria Oliver and her team, ValenciaIA4COVID, have been declared the world’s winners of the $US500,000 X Prize, which saw them create models for 236 countries to help navigate COVID-19.

“The models can make decisions that are not based on political interests or instinct but on the evidence of what the data is telling you,” Dr Oliver says.

“The system proposes policies with the best cost-benefit tradeoff so if your hospitals are empty, then you might want to open up a bit or if the hospitals are full, you would probably need to implement stricter interventions to contain transmission.”

Dr Oliver’s complex models take into account different measures but the final challenge to win the prize was to make an “artificial politician” - also called the prescriptor - to recommend an accurate response across 236 different countries or regions.

“Australia was very easy - you predict zero cases and you get it right,” she quips. “This is the advantage Australia and New Zealand have of being an island. But you will also have to vaccinate so you don’t get left behind.”

This video summarises the Xprize competition and methods.

The maths is complex, the outcomes were not

The coronavirus pandemic created plenty of firsts - it was not only the first time the world adapted ZOOM meetings faster than you could say ‘have you washed your hands’ but also the first global pandemic that captured and shared daily COVID-19 case data across the globe.

Dr Oliver was working with the President of Spain’s Valencia region in December 2020 when the team started work on the X prize competition, and believes she couldn’t have done the work without the daily collaboration she was having at the highest levels of government.

Building an accurate prediction for COVID-19 cases came using official COVID-19 case data and the Oxford COVID-19 Government Response Tracker data set as the main data sources.

The team built neural network-based computational epidemiological models to predict COVID-19 cases 30, 60 and 180 days into the future.

The models used the confirmed number of COVID-19 cases and the implemented interventions to contain the pandemic in each of the 236 countries or regions in the world since March 2020.

The team also had to create a ‘prescriptor’ of 12 possible interventions, which had three to five different ‘levels’ each. These interventions included:

  • Schools closing

  • Workplaces closing

  • Cancelling public events

  • Restricting public gatherings

  • Closing public transport

  • Requiring people to say home

  • Restricting internal movement

  • International travel controls

  • Public information campaigns like education

  • Testing policies

  • Contact tracing

  • Facial coverings, or mask-wearing

Like all good data science, the quality of data varied so some countries had to be excluded due to lack of reliable data. The details of the work are in the process of being published in scientific journals but is not yet available, though there is more detail on this website.

Dr Oliver says workplaces closing so people can work from home were the most effective policies driving the number of COVID-19 cases up or down across all 236 countries and regions, closely followed by education interventions, international travel controls and restrictions on gatherings.

Better modelling came as data augmented with survey data

The team were able to refine their models as Spain embarked on its third wave of COVID-19, raising the death toll in Spain to around 77,000 deaths in April 2021 - just as they finished the predictor model for X Prize.

“We were able to accurately predict the evolution of the third wave of infections and test our predictions against real data,” Dr Oliver says.

Dr Oliver has also been running one of the world’s largest COVID response surveys, gaining valuable qualitative insights into the pandemic and offering citizens a voice.

This survey - which has had 600,000 responses - revealed valuable information about the perception and impact of the pandemic on people’s lives, especially on young people as well as the limitations of contact tracing.

“According to our survey, the contact tracing apps are a complete failure in Spain, Italy and Germany,” she says.

But how accurate was the AI-built perfect politician?

Given the hypothetical nature of the ‘prescriptor’ - or the so-called AI politician - the team could not evaluate its performance against ground truth.

The predictor did, however, deliver policy recommendations in less than two hours, which shows that politicians and governments could be using these types of models to help navigate complex daily decision making.

“It’s my hope that it will also help build the case for more evidence-based policy decision making,” Dr Oliver says.

“Public administration and governments of all levels are not undergoing the digital transformation that most companies have had to do in the past 15 years and the need to put in place a way to systematically gather data to inform policies in the future.”

Her greatest hope is that governments start standardising data collection and systematic analysis for the social good.

“We can easily create a virtuous cycle between data, people and technology so public policy is informed by insights,” she says.

Security by design: avoiding cybersecurity pitfalls

Cybercrime is on the rise, with small and medium businesses without resources to maintain data security targeted for data theft, phishing scams and personal information.

Cyber criminals are agile, smart and find it easy to breach data security – they buy kits on the dark web to give them step-by-step instructions on how to hit businesses with a Denial of Service attack for $US6, or use bitcoin to buy compromised emails, passwords and even credit card details.

It’s a cinch for crims to launch phishing attacks to steal passwords and personal information. The Internet of Things - our voice-activated speakers or fitness trackers - are also a “threat vector” for security breaches.

“They target small to medium enterprises and just this year, the confirmed losses reported is over $160m,” says Australian Federal Police Superintendent Mark Cobran, who heads up cybercrime operations in Canberra.

With last month’s Nielsen monthly television ratings held up by a cyber attack and our Prime Minister investing money to prevent rising cybercrime attacks, the risk of being a victim of poor data security is higher now that we work remotely and on the Cloud, or tether to our mobile device for internet access or use a compromised VPN.

“Criminals are motivated by the profit motive – they are trying to steal money or steal information that they can sell. That might be personal details or it could be proprietary details from a business, or even an email contact list or anything else they can sell on the darknet to make money,” Cobran explains.

The COVID-19 pandemic has seen a rise in cybercrime, with hackers exploiting known vulnerabilities in work-from-home tools like Citrix or setting up elaborate scams that look like legitimate government websites or stealing personal information to divert people’s $10,000 superannuation payments into their own accounts.

“They are very agile and use techniques such as contact tracing - for example people receive an email that looks legitimate from the Department of Health saying you’ve been reported as someone who’s been in contact with someone who is COVID positive – or they say people are eligible for a rebate and they ask people to click a link. They use spook websites to look legitimate,” Cobran says.

So how can businesses secure their data?

“The general public needs to become more vigilant,” says Blue Bricks CEO Vikram Sareen, who offers an ethical hacking service to help businesses secure their systems. Startups and small businesses also need to care about data security more than they currently do.

With no shortage of criminals trying to deny, destroy or disrupt businesses - especially small to medium sized businesses holding valuable contact lists or intellectual property on their systems - the Essential Eight is one simple technical guide to data security.

The Essential Eight outlines technical data security steps - ranging from application control through to multi-factor authentication and daily back ups - to mitigate cybercrime risks.

Security experts like Avertro CEO Ian Yip also say that businesses should implement a culture of security and train their staff. Though Vikram Sareen points out that businesses can’t afford to make that training complex.

“Most of the training is extremely boring - it’s a huge amount of complex information downloaded on people. If there is a rewards-based awareness or learning or gamification of how to teach security, then it will make more sense,” Sareen says.

AFP Superintendent Cobran agrees that training is key.

“Investigations with major multinational companies who have had data stolen have been caused by a lack of training or lack of rigorous enforcement around what people should do,” he says.

Senior principal research scientist at CSIRO Data 61 Dr Surya Nepal says workplaces need to be talking about security at every board meeting - and not just because the financial regulators have now made cybersecurity a board responsibility.

“At work, we start a meeting by acknowledging our traditional landowners and then we should be talking about data security. It must be a part of every meeting,” he says, arguing that building privacy and security by design quickly becomes outdated as technology moves on.

“Don’t collect the information you don’t really need. Do you really need a phone number or a credit card details? The more data you collect, the more you make yourself a target.”

Creating a data security culture

Building privacy and security into a workplace culture not only demands training, but also a documented procedure around what people need to do if a breach occurs.

Hall Chadwick partner David Watt - a forensic accountant who helps quantify the costs of cyberattacks - says key staff need a documented Data Breach Response Plan to know what should be done and in what order.

“It might be a good idea to have that plan written in hard copy somewhere – when faced with ransomware, your systems will be down and you may be isolated from the world of the internet for a period of time until you can get back up and running,” he says.

Once the breach has happened, it’s also important for businesses to review the incident and modify procedures and policies to keep improving their response.

Watt also says each cyber breach tends to cost a business around $250,000 per event - so taking preventative measures is vital. Some businesses may see data security as a black hole for money, but it’s worth investing in to protect your assets, especially data assets.

“You’re under a legal obligation to protect personal data in this country and if there’s evidence to support that you don’t do that there can be fines for the company and the individuals involved,” Watt says.

AI-ifying customer experience

Originally published in Information Age

From columns and rows to collaboration and compliance, data-led insights and new technology is enabling a brave new world of data-led marketing opportunities.

Data is a four-letter word that can overwhelm or excite a business in equal measure.

The deluge of details collected by most modern businesses on transactions, customers, sales and even electricity usage has changed the way many businesses work, upending how technology and customer marketing work together.

Data and cloud-based technology offer the capability to automate, personalise, create efficiencies and engineer better customer outcomes than any analytics or data expert who remembers floppy disks could ever have imagined.

Yet data and technology-enabled businesses also risk higher costs, complex compliance demands and must constantly adapt to new business practices like greater collaboration and better staff attraction policies to retain skills that are in high demand.

“You don’t understand if you have a good data scientist until you’ve spent millions of dollars and got it wrong - it can be an expensive mistake,” said Peter Bonney, general manager of Technology, Engineering and Data at grocery giant Coles.

“People have to remember that technology, data and things like artificial intelligence are only tools to solve business problems, not solutions in and of themselves.”

On the flip side, head of marketing at telecommunications giant Optus - Melissa Hopkins - said there was an abundance of new companies, vendors and technologies enabling new opportunities to connect with customers and deliver business value in brave new ways.

“We are at an intense point in time that is very exciting. We need to rewrite the rules to take advantage of the opportunities,” she says.

Data informed decision-making and marketing

Liquid CX founder Simone Blakers - who has worked at the bleeding edge of data marketing for two decades - says the rise of ‘plug and play technology’, or SaaS products and APIs, has opened the door for all levels of business to take advantage of automation and personalisation.

“It’s now easy to combine sales data, online data and customer data to a highly sophisticated level that can be accessed by all sizes of business,” she said.

Larger companies still have greater advantage, as they can quickly leverage resources to explore new technologies and gain commercial advantage.

Blakers has seen big companies leverage artificial intelligence technology to test and predict the impact of communications, packaging and simulated retail experiences. The technology enables this testing to be done in real time and at scale before going to market, offering large companies the ability to optimise their customer experience faster than ever before.

“Using predictive and real time analytics on the fly changes the very nature of creative and retail marketing, as ideas are tested before they are even made or put into market,” Blakers said.

“There is now business intelligence technology with voice assistant interfaces that let you pivot, understand triggers quickly and analyse a campaign – all through one verbal query.”

“The secret world of data intelligence is now democratised,” Blakers said. “You can, however, get lost in a sea of data if you don’t manage it well.”

Hopkins said corporations have gone from relying on large partners like software vendors or agencies to working with more nimble partners who can allow big business to ‘dip their toe’ into new technology offerings.

“The small indies that are popping up have leaner teams. Technology partners like Facebook and Google and Salesforce remain important but so do the smaller AI companies that specialise in particular areas - smaller organisations have less fat where you deal with the thinkers and doers,” she said.

As an example, Optus worked with a supplier to place a media value on the telecommunications giants owned assets - everything from the screens inside Optus retail stores to the company’s email list, website traffic, app usage and more was valued as a media asset for the business.

There are an abundance of clever startups leveraging new ML and AI technologies to partner with businesses to create better insights and better customer outcomes.

Blakers cites and as two interesting examples but predicts that technology offering the greatest business impact will be most in demand. “The value always lies in the use case, not the technology itself,” she said.

Leveraging artificial intelligence and machine learning for prediction

Collecting data may be easier than it used to be. But working with data has rarely been.

Merging data sets, cleaning data and finding the right insights is as hard as it was before technology enabled more opportunities.

“There is a small grey patch of hair on my head for data quality issues,” Peter Bonney said.

What’s more, the business use cases for machine learning and artificial intelligence may be exciting, but it can be complex to navigate.

As Melissa Hopkins said: ““Yes, AI is important for marketers and businesses but you can’t take your hands off it. Unless you know how to drive it, stay away from it or leave it to the experts.”

Algorithms and models can degrade over time, or become entirely inaccurate if a brand new team member in another department isn’t keying in data correctly.

Blakers said business leaders needed to understand the role artificial intelligence and machine learning could play in their company strategy.

“Investing in innovation is not just about the ROI, it is also the COI – cost of inaction. You don’t want to wake up and realise your competitor has spent the last 5 years training AI to better service their customers and now you can’t possibly catch up,” she said.

“Executive leaders who aren’t data experts and rely on the data geeks to tell them what’s happening need to know the right questions to ask and break through the complexity.”

Collaborating across the business

How organisations solve business problems with data and technology is more nuanced than “just do it”. Leveraging plug and play packaged applications can turn an organisation into a fast bowler, but that doesn’t always win the long game.

“If you’re a company that relies on (technology) partners and SaaS products, you will have to build data and technology capabilities in house and it will be a longer journey for you,” Peter Bonney said.

“If you do have the technology and data discipline in an organisation, then you can build analytical products for fast-changing areas. The real competitive advantage is how quickly you can get things in market.”

Bonney said large and complex businesses needed business managers to partner with data and technology teams to manage projects more like a product than a one-off implementation.

“Businesses need senior data scientists with a product development mindset - it’s about a hypothesis, test it, test the outcome, measure it and rinse and repeat,” he said.

“In the old days, technology teams would hold a governance forum that the rest of the business didn’t understand or care about. But now it’s about what’s broken, how can we fix it and how can we get more insights to deliver a better experience for the customer.”

Hopkins said human insight will drive the best data and technology strategies. “If you want to use AI, make sure there is input from humans. It’s human insight that will drive the 360 degree customer experience,” she said.

She said the Optus marketing team have in-housed functions like SEO and SEM but their sophisticated digital team blends with other business functions rather than solely focusing on ‘digital’ or ‘data’.

The job spec for marketing has to change and it is changing. We all have to innovate and do more than merely use different channels and models.

Machine learning means businesses can now personalise communications and marketing and data drives more decisions across the business value chain.

“But data is like crude oil - it’s valuable but worthless unless it can be refined to come with insights,” Hopkins said.

The skill shortages in data and technology

Fast and furious change across the data marketing landscape inevitably leads to skill shortages, which can make it hard for all businesses to adapt to the new opportunities.

Blakers says the strategic lens and skillset is more necessary than ever, as specialists have evolved at the channel or platform level but not necessarily understanding the complexity of the business landscape.

For Bonney at Coles - which faces rising disruptive threats from global data giants like Amazon - the shortage is in cloud engineer specialists and data scientists with deep market experience in areas like forecasting or personalisation.

Compliance issues - especially around security, privacy and personal information collection - were forcing senior executive leaders to understand data and technology issues more broadly than they ever needed to in the past.

“Skills have always been challenging, but ultimately you don’t want anyone with their head in a spreadsheet without having their eye on the customer,” Bonney said.

Data science vs data analytics

Phrases like artificial intelligence, data analytics, data science and data visualisation are buzzwords that blend into a soup of similarity - here's a simple explanation.

The jargon of data-centric marketing can sometimes be enough to set your teeth on edge and activate eyerolls, but defining data science versus data visualisation and the differences between business analytics and statistics is worth taking the time to do.

(Ahem, especially before we embark on the complexity that sits beneath the umbrella of artificial intelligence and the amazing tools it can unleash for business and marketing.)

Data science relates to research, intelligence, modelling and analytics but it is worth understanding why it is different to the other terms - and not just because data scientists get paid more than business analysts or researchers.

Data science lecturer at University of Technology Alex Scriven says analytics and data science overlap in terms of exploratory tools, techniques and data visualisation - and he has just written a book called From Business Intelligence to Data Science (Manning Publications) which outlines this in detail.

The simple way to sum up data science versus data analytics is that analytics looks backwards while data science tends to look forwards.

“Analytics is still answering questions of your data, such as 'what kind of products do consumers buy?' whereas data science may ask a question like 'can you predict what this customer will buy?',” Scriven says. “Even something like correlation matrices and some hypothesis testing could be considered analytics, which also has a place in data science.” So it does cross over and tends to get confusing.

NSW Government Chief Data Scientist Ian Oppermann says the opportunity for data science exists when you have an impossible question. The crunching of large datasets to build models or predict outcomes using data science is powerful.

“Data science is good at building predictive models and looking at root cause and ‘what if’ scenarios. Statistics and analytics looks at things differently,” he says, using the example of a model he built to prevent fire and rescue teams wasting time attending false fire alarms.

Statistics and analytics showed that 97% of fire alarm callouts were people burning toast rather than a genuine fire. Yet data science enabled a team to build a predictor model to try to determine when each alarm would be false or genuine.

The data scientists aggregated social media data, weather data, lunar cycles, pollen counts and fire alarm panel data to predict when the fire alarm callouts would be genuine.

“The model predicted with 77% accuracy, with the lunar cycles accounting for 2.5% of that accuracy,” Oppermann says. “It made a sceptical fire commissioner learn to love what we could do with data science.” Analytics, by comparison, typically only allows you to look at the data and draw out metrics that exist in it.

Data means nothing if you can’t see the meaning

Technical tools like Tableau and Power BI are powering business enterprise visualisations that go beyond the average Excel chart, making business enterprise reporting way more interesting than it used to be.

Savvy business leaders can move beyond a stagnant monthly sales report graph to see their own business dashboards updated in real time - you can even find beautiful examples on .

The rise of data visualisation goes beyond business reporting. Take a look at David McCandless to see how big data is changing media and storytelling.

Data visualisers like Priya Ramakrishnan - who worked in - says the real skill of data visualisation comes from being able to clean and structure the data properly before analysing it to find the story hiding in all the numbers.

So what about Artificial Intelligence

If there is one little term - AI - hiding big opportunities, it is artificial intelligence, which is allowing a multitude of business processes to be automated in new and efficient ways.

Machine learning (also called ML) is a branch of artificial intelligence, but typically the terms all get used to refer to the same thing.

“Recommendation engines, customer churn models, sales forecasting, machinery maintenance are all areas that classic ML can be used (across business),” says Scriven.

CSIRO Data 61 Senior Principal Researcher Dr Surya Nepal believes AI is where we will see many Australian businesses advance quickly to automate and solve bigger business problems.

AI also refers to deep learning, a powerful form of machine learning based on mathematical operators that simulate the structure of the human brain and can make amazing predictions when you have large quantities of (reasonably) structured data.

Sales forecasting, attribution modelling and lead scoring can make more powerful predictions with deep learning than analytics or statistics alone. There is a catch, though as many neural nets don’t actually let you know the precise factor or cause – it often just “is” or “isn’t”.

The Centre for the Future founder Richard Hames says the 2020s will be a perfect storm of disruption with the rate of apps, tech platforms and knowledge channels continuing to accelerate.

“What marketers need to do is stretch their imaginative capabilities and embrace what’s possible,” he says.

Data protection laws in australia

Australian data protection laws evolve as new public bodies, frameworks and notification laws come into place with the arrival of changes like the Consumer Data Right.

Australian data protection laws have evolved quickly to regulate business data, which was once hyped by The Economist as the ‘new oil’. The World Economic Forum delivered a reality check, pointing out data is more like gold than oil - it must be mined before it’s useful and is best kept away from criminals.

Australia’s regulators and legislators seem to agree with the World Economic Forum, creating a raft of legislation, new regulatory bodies and frameworks to deal with the fast-changing technology landscape.

“Regulation is very much behind the art of the possible, when it comes to technology,” says NSW Government Chief Data Scientist Dr Ian Oppermann.

“We need to think carefully about the use of data, cybersecurity and privacy. The real question is what is the risk framework that needs to be in place so we can reach utopia rather than dystopia.”

The arrival of the Consumer Data Right (CDR) in Australia on July 1, 2020, is inextricably linked to privacy and cybercrime prevention regulation in Australia, prompting new government rules. Most of the rules and regulations rely on notifying regulators of breaches.

The CDR allows consumers to share their personal banking data - which used to be seen as the property of the banks - to land a better loan or credit card deal, creating impetus for financial organisations to invest in data security and privacy practices. Businesses now need to act as trusted custodians of people’s data, rather than exploiting the riches such data can bring.

Non-technical business stakeholders may once have shut their eyes to the costs and complexity of data handling, but the consequences are now too high if privacy or security go wrong. Just ask Landmark White, who estimated they lost $50m after being crippled by a rogue software contractor’s misappropriation of property valuation data.

With Europe’s strict GDPR laws paving the way for Australia’s CDR, there’s no doubt that privacy, data security and rising cybercrime are firmly on boardroom - and government - agendas. Many large companies are investing to voluntarily comply with IT security standard ISO 27001, particularly since APRA made it clear that board members can be liable for security and privacy risks.

The worry is that small and medium-sized businesses could be left behind and become bigger targets for criminals because they don’t invest in the privacy protections and security they need. So where can we expect the data protection laws to head?

Data protection landscape will shift with privacy and security issues

The Federal Government has made it clear that protecting the economy from cybercrime will be a priority, announcing more funding to invest in technical capabilities.

“The regulatory landscape for cybercrime and privacy will continue to evolve with encryption laws, privacy laws moving closer to Europe’s GDPR and the laws that apply to cybercrime continuing to evolve,” explains Dr Surya Nepal, CSIRO Data 61’s senior principal research scientist.

Cybersecurity experts like Blue Bricks CEO Vikram Sareen say Australia is still well behind other countries around the world, but there will be a big bubble of investment in security over the short term.

Privacy Impact Assessments are a business tool that’s fast becoming the new black as companies rush to protect themselves from growing crime and privacy risks. CSIRO Data 61’s Dr Nepal says the most important thing is for businesses to invest in protecting their core data and intellectual property.

“Define your core data - this might be your business IP, your stakeholder information ... it’s not only customers but also suppliers and your financial information. Even big businesses have a lot of trouble managing this,” he says, arguing businesses must regularly attend to ‘cyber hygiene’ and check who has administrative access to what data.

Dr Nepal hopes AI developments help unleash an affordable “cybersecurity in a box” modem-stye solution that small to medium businesses can plug into their network to monitor threats and data leaks.

“At CSIRO, we think something like this could be a usable security solution that can analyze and alert you to threats in the same way a home security system can alert you to threats, but not necessarily protect you from them,” he says. It will also be more affordable than hiring six-figure security experts to consult for a business.

Government will try to stay ahead of the game

Governments will keep trying to regulate as technology shifts and evolves, but often have to take the time to build consensus and write long reports before they can take action or commit funds.

When it comes to data handling and privacy issues, Dr Ian Oppermann says the Five Safes Framework developed by the privacy regulatory OAIC is helpful for handling data, and ultimately giving data back to the people who create it. After all, the general public cannot walk down a city street or catch public transport without the government capturing this data on our behalf.

“Smart Cities” are a new buzzword which allows governments and technology companies to partner and will start seeing new technologies across cities. Governments are also likely to continue to create digital-centric brands like “Services Australia” as they attempt to use digital technology to scale services.

Dr Nepal says regulation of the Internet of Things - smart TVs and driverless cars - will be the area to watch.

“The television you buy has energy ratings and the food we buy has health star ratings, but that Smart TV needs a security rating,” he says, pointing out that there is more regulation in buying a child’s toy than buying internet-connected devices.

“Security ratings would be a good direction to go in, but there is a risk it might stifle innovation.”

And there lies the rub - regulating can kill the very benefit that data and technology can bring before it’s even had a chance to be made real.