For the past few months, an independent board of technology experts has been closely tracking the new ways that AI and data have been used to counter and mitigate the effects of the COVID-19 pandemic in the UK; and now, they are lifting the veil on the good, the bad and the ugly of the past year in digital tech.
The Center for Data Ethics and Innovation (CDEI) has released a new report diving deep into the 118 individual use-cases for AI and data-driven technologies that have been added to the organization's COVID-19 repository since last November.
Spanning vastly different sectors and locations, the examples collated in the document provide a unique vision of the ways that technology can help in a time of crisis. From piloting drones to delivering medical supplies, to monitoring the behavior of residents in public transport during the easing of lockdown restrictions: if there is one observation that all experts will agree on, it is certainly that technology has been a central pillar in the support of the response to the pandemic.
"While public attention largely centred on high-profile applications aimed at either suppressing the virus or coping with its effects, our research highlights the breadth of applications beyond these two use-cases," says the report.
"Data-driven technology and AI has been used for a multitude of purposes in the fight against the virus. From connecting volunteers on social media platforms, to identifying treatments and vaccinations for COVID-19, almost every facet of the response has required the support of data."
SEE: COVID-19: A guide and checklist for restarting your business (TechRepublic Premium)
Key to the development of new technologies was an increased sharing of data – whether for small, local schemes, or for widespread global initiatives. With new data-sharing agreements in place, for example, a platform was set up by the NHS to track infections and death rates at the national level; similarly, the Greater London Authority (GLA) created a COVID-19 API to ensure that data on the virus was consistent. Meanwhile, opening up databases meant hospitals were able to see and share where equipment was free and where it was needed.
Data analytics also enabled a better delivery of public services indirectly related to the medical field, especially in local councils. The CDEI pointed to the example of Hackney Council in London, which joined up various datasets for the first time in a single platform to gather a live view of the most clinically and economically vulnerable residents in the area. The council used residents' unique property reference numbers (UPRN), and aggregated those with data such as housing benefit, council tax, adult social care or children's and families' services. This built up an informed picture of each individual property in Hackney, helping the council target those most in need of help.
Artificial intelligence, although not used to the same extent as less advanced data analytics tools, still played a role in easing the impact of the crisis. Aside from assisting the momentous task of developing a vaccine against COVID-19, AI was used to drive models for rapid virus detection. Researchers at Oxford University Hospitals, for example, built an AI-driven test that could screen patients for COVID-19 with 92% accuracy within an hour of their arrival.
AI systems are also widely integrated into technologies that might help with the easing of lockdown restrictions. Some companies – including marathon organizers – are banking on wearable devices that could notify their staff when they are in close proximities to others. Smart city technologies monitoring pedestrian flows, traffic or air quality were used to help authorities understand the extent to which people were staying at home, and where they were headed when they were outdoors.
Behavior AI platform developer Humanising Autonomy partnered with Transport for Greater Manchester (TfGM), to provide insights from CCTV video footage at busy sections in the public transport system. Anonymized content was processed by AI software to understand how passengers were changing their behavior because of the pandemic, and better adapt infrastructure and timetables to improve services.
But despite the impressive effort from public services to come up with innovative solutions to tackle the challenges of a global health crisis, the CDEI stressed that much more, and much better, could have been achieved. Forced to urgently share data to function more efficiently, government departments and councils rapidly put in place the necessary legal agreements, technical standards and oversight measures; but in many cases, those key foundations that enable the use of data were not in place already.
"Data storage and data sharing throughout the pandemic has, of course, not been without its challenges," reads the report. "Even though we have seen many positive examples of these practices, there are still lessons to be learnt, and many barriers to the more effective use of data continue to endure."
In the public sector, the deficiencies are especially striking, and the issue has been the main topic of many reports already. Only a few months ago, the CDEI published an analysis of data sharing between government departments and the private sector, and found that fundamental legal, technical and cultural barriers still stubbornly remain in the way of data-informed innovation.
More recently, a report from independent think tank the Institute for Government (IfG) described a cocktail of legacy IT, poor coordination and lack of skills still pervading public services' use of data. The IfG called for the government to "fix the plumbing" when it comes to the management and digitization of information, suggesting that the problems are deep-rooted and require a complete re-think.
So, what's the verdict?
While it is safe to say that the impact of COVID-19 would have been even greater without the contributions of technology, the CDEI argued that it is too early to properly evaluate the impact of the innovations registered in the repository.
It also surveyed 12,000 individuals during six months; almost three-quarters felt that digital technology had the potential to be used in response to the outbreak.
When it comes to whether or not the technology is actually working, responses were more nuanced. More than a third of the public (37%) blamed the potential failings of digital solutions on the technology itself – either anticipating that the right technology wouldn't launch in time, or that it wouldn't work. At the same time, less than 42% of respondents said that digital technology was making the situation in the UK better.
This suggests an opportunity gap: while the public is confident that tech can help, citizens are yet to be convinced that it already is.
The past year has seen a number of high-profile fiascos when it comes to the use of data-informed technologies. For example, last summer, the exam regulator Ofqual used a biased algorithm to predict A-level and GCSE results, which caused student outrage across the country as grades were allocated in a discriminatory way. Around the same time, the NHS finally released a contact-tracing app – but only after several months of delays, marked by significant technical glitches and U-turns on original plans.
These failings have only contributed to the undermining of public trust in the use of digital technologies; and yet public trust is central to the successful deployment of data-driven tools.
The CDEI's poll, in fact, revealed that trust in the rules and regulations governing technology is the single biggest predictor of whether someone believes that digital technology has a role to play in the COVID-19 response. Age, education, social media usage: all played a smaller role in determining individuals' level of support for data-driven innovation.
The report, therefore, seems to draw a clear conclusion: better data frameworks and governance are the only way to close the opportunity gap with useful AI and data-driven technology. The journey to securing the public's trust will be long-winded and challenging.