“Technology is never a silver bullet and it’s particularly not a silver bullet in this circumstance,” Carly Kind, director of AI research center Ada Lovelace Institute, told POLITICO. “But it’s often being framed as a silver bullet — and that’s part of the problem.”
Her warning comes as European governments from Berlin to Valletta are racing to roll out contact-tracing apps that analyze signals between mobile phones to alert people who have been close enough to infect each other.
Widespread downloading of such apps, they hope, will help to track the virus’ spread and ease the lifting of lockdowns.
But while such plans have prompted a fiery discussion among Europe’s privacy regulators, little attention has centered on the effectiveness of apps in stopping the pandemic.
Tech experts harbor serious doubts.
In an analysis released this week, the London-based Ada Lovelace Institute warned about what it described as a lack of evidence that apps help to curb the spread of the virus, urging the government to shelve plans for a U.K. contact tracing app.
Researchers across Europe are striking similar notes, warning that apps alone will not on their own allow countries to fully resume public life. That stage will only be reached when a vaccine is deployed, they say.
Back to square one
“Using a track-and-tracing app might be something that helps, but it’s not — as a lot of people now think — a catch-all solution that allows you to lift all the restrictions and we’re done,” said Frank Dignum, a professor of socially-aware artificial intelligence at Umeå University in Sweden.
Along with a team of two dozen AI researchers, Dignum has built a computer model that can simulate the consequences of policy measures to curb the coronavirus.
When they used the model to test the effectiveness of contact-tracing apps, the results suggested that they are far less effective than often claimed.
For example, they found out that it will require far more than 60 percent of the population to download the app — an oft-quoted magical threshold — in order for them to be effective. A key reason, they said, is that most of the people downloading the app will be people avoiding infection rather than those at particular risk of contracting the virus.
Any rollout of contact-tracing apps would also have to be paired with mass testing.
“Using a track-and-tracing app might be something that helps, but it’s not — as a lot of people now think — a catch-all solution that allows you to lift all the restrictions and we’re done,” said Frank Dignum, a professor of socially-aware artificial intelligence at Umeå University in Sweden.
Along with a team of two dozen AI researchers, Dignum has built a computer model that can simulate the consequences of policy measures to curb the coronavirus.
When they used the model to test the effectiveness of contact-tracing apps, the results suggested that they are far less effective than often claimed.
For example, they found out that it will require far more than 60 percent of the population to download the app — an oft-quoted magical threshold — in order for them to be effective. A key reason, they said, is that most of the people downloading the app will be people avoiding infection rather than those at particular risk of contracting the virus.
Any rollout of contact-tracing apps would also have to be paired with mass testing.
When Dignum’s team ran the simulation for his home country, the Netherlands, he noticed that a contact-tracing app scheme would have to be paired with more than 100,000 daily tests — far more than the current number. Failing that, all those who have potentially been in contact with infected individuals would have to go back into self-isolation.
“Within one or two weeks, you would have one to two million people sitting at home again,” he said.
This prompted Dignum last week to co-sign an open letter along with 130 other researchers from the Netherlands urging Prime Minister Mark Rutte’s government to assess how useful and effective such apps are before rolling them out.
A potentially deadly dilemma
After countries across Europe imposed unprecedented measures to restrict movement, statistics suggest that lockdowns have helped to flatten the curve of new infections.
The good news, however, means that decision-makers face a tough dilemma over what to do next.
While they are under pressure to lift at least parts of the lockdowns due to economic and other concerns, they worry this will lead to another wave of new infections with potentially even more devastating consequences.
To prevent that, decision-makers from the U.K. to France have increasingly been promoting the use of digital tools, ranging from „immunity passports,“ proving that someone has immunity from contracting the virus, to e-health applications allowing patients to assess their symptoms remotely.
But no technology has gained as much traction as contact-tracing apps, which are supposed to help countries prevent large-scale outbreaks before they happen.
The experience of countries such as Taiwan or South Korea suggest that such measures may be effective — if paired with mass testing.
Kind warned that „people try to boil that success down to the use of one a single technological measure such as contact-tracing — when, in fact, technology … was embedded in a set of policy and strategic decisions such as widespread testing or temperature-taking.“
„And you can’t separate the success of the technology from the policies,” she added.
Experts also warn that the underlying technology, which analyzes Bluetooth signals, is prone to errors: In cities, for example, next-door neighbors who have never met could end up being alerted about each other’s infections even if walls between them make contracting the virus impossible.
Such “false positives,” particularly once their number grows, could make people increasingly indifferent to being alerted about potential contacts with others who are infected.
Finally, it is far from certain that European governments can convince enough citizens to use the voluntary apps. While surveys from places like the U.K. suggest that a majority of citizens support the schemes, people in other countries such as Germany are more skeptical.
And even in Singapore, where three out of four citizens say they trust their government with their data, only about 1.1 million of 5.4 million citizens have so far signed up for the country’s voluntary scheme — far below a 75 percent threshold authorities in the country have said is necessary for it to be effective.
Source: politico.com
See more here: news365.stream
