Prediction

21 Jul 2020
The algorithm and mathematical model used to predict students’ grades by the International Baccalaureate programme, which was forced to cancel exams because of the pandemic, incorporated three elements: coursework, teachers’ predictions of their students’ exam grades, and “school context”, which was
09 Jun 2020
Data-driven companies like Experian, CACI, and Xantura are pitching their services to help UK local officials to identify people in need. Xantura and CIPFA, the accountancy body for the public sector, have teamed up to deploy a £15,000 tool that uses local authority data including the NHS’s
11 Jun 2020
Nearly six months after the emergence of the coronavirus, only 7.1% of research on COVID-19 references AI compared to 12% of research on other topics. AI is being used to make predictive analyses of patient data, especially medical scans, and analyse social media data, predict the spread of the
30 Apr 2020
Sweden's Lund University has launched an app intended to map the spread of the coronavirus across Sweden, a localised version of the JoinZOE COVID Symptom Tracker app pioneered in the UK, which the researchers believe could be coupled with seroprovalence testing in order to develop an accurate map
In September 2018, at least five local English councils had developed or implemented a predictive analytics system incorporating the data of at least 377,000 people with the intention of preventing child abuse. Advocates of these systems argue that they help councils struggling under budget cuts to
By 2018, the Danish municipality of Gladsaxe, in Copenhagen, began identifying children at risk of abuse so that flagged families could be targeted for early intervention by applying a set of specially designed algorithms to information already gathered by the centralised Udbetaling Danmark
In November 2018, worried American parents wishing to check out prospective babysitters and dissatisfied with criminal background checks began paying $24.99 for a scan from the online service Predictim, which claimed to use "advanced artificial intelligence" to offer an automated risk rating
A study published in January 2019 found that a form of facial recognition technology that interprets emotions in facial expressions assigns more negative emotions to black men's faces than white men's faces. The problem is the latest in a series of ways that facial recognition has failed for non
In 2018, technology companies and medical providers were experimenting with machine learning and AI to mine health records and online posts to identify patterns linked to suicide, hoping to be able to predict, and therefore prevent, such attempts. On the academic side, a pilot programme conducted by
16 Nov 2017
In 2017, US Immigration & Customs Enforcement (ICE) announced that it would seek to use artificial intelligence to automatically evaluate the probability of a prospective immigrant “becoming a positively contributing member of society.” In a letter to acting Department of Homeland Security Secretary
30 Apr 2019
The New York Times picked 16 categories (like registered Democrats or people trying to lose weight) and targeted ads at people in them. They used the ads to reveal the invisible information itself, noting that it is a "story of how our information is used not just to target us but to manipulate
15 Aug 2018
In 2018, Wells Fargo disclosed that due to a computer bug that remained undiscovered for nearly five years 600 customers were granted more expensive mortgage loans than they could have qualified for. About 400 of them went on to lose their homes. The announcement reignited the public anger and
09 Jul 2018
In 2018, economists Marianne Bertrand and Emir Kamenica at the University of Chicago Booth School of Business showed that national divisions are so entrenched that details of what Americans buy, do, and watch can be used to predict, sometimes with more than 90% accuracy, their politics, race, income
08 May 2018
In 2018, based on an analysis of 270,000 purchases between October 2015 and December 2016 on a German ecommerce site that sells furniture on credit, researchers at the National Bureau of Economic Research found that variables such as the type of device could be used to estimate the likelihood that a
12 Jun 2018
In June 2018, Uber filed a US patent application for technology intended to help the company identify drunk riders by comparing data from new ride requests to past requests made by the same user. Conclusions drawn from data such as the number of typos or the angle at which the rider is holding the
02 Apr 2018
By 2018, gene studies involving more than 200,000 test takers had found correlations between 500 human genes and academic success. Based on these results, the behavioural geneticist Robert Plomin claimed that parents would be able to use consumer tests to enable "precision education", built around
17 Sep 2018
In 2018, at least five British local authorities began developing systems intended to use predictive analytics to identify families needing attention from child services on the basis that algorithmic profiling could help them target their scarce resources more efficiently. Data about at least 377
The accuracy of Facebook's ad targeting sometimes leads users to believe that Facebook is spying on them by tapping the microphones in their phones. Facebook has denied the practice - and is likely telling the truth because uploading and scanning the amount of audio data such a system would involve
In 2015, the University of Arizona began tracking freshman students’ ID card swipes as part of a project to try to lower the rate at which students drop out or leave for another university. The cards, which include an embedded sensor and are given to all students, can be read at almost 700 locations
Designed for use by border guards, Unisys' LineSight software uses advanced data analytics and machine learning to help border guards decide whether to inspect travellers more closely before admitting them into their country. Unisys says the software assesses each traveller's risk beginning with the
Car companies have long collected data about the consumers who buy their cars. Now, they hope to aggregate and sell customer preferences to outside vendors for marketing purposes much as online tech giants like Google and Facebook already do. The companies say that exploiting this data will help
In 2016, Facebook and its photo-sharing subsidiary Instagram rolled out a new reporting tool that lets users anonymously flag posts that suggest friends are threatening self-harm or suicide. The act of flagging the post triggers a message from Instagram to the user in question offering support
In a study of COMPAS, an algorithmic tool used in the US criminal justice system , Dartmouth College researchers Julia Dressel and Hany Farid found that the algorithm did no better than volunteers recruited via a crowdsourcing site. COMPAS, a proprietary risk assessment algorithm developed by
02 Jan 2018
In February 2018 the Canadian government announced a three-month pilot partnership with the artificial intelligence company Advanced Symbolics to monitor social media posts with a view to predicting rises in regional suicide risk. Advanced Symbolics will look for trends by analysing posts from 160
In 2014, the UK suicide prevention group The Samaritans launched Radar, a Twitter-based service intended to leverage the social graph to identify people showing signs of suicidal intent on social media and alert their friends to reach out to offer them help. The app was quickly taken offline after
Among the friends Facebook recommended to Kashmir Hill as people she might know was Rebecca Porter, to the best of her knowledge a total stranger. Because Hill was studying how the "black box" of Facebook recommendations worked, she contacted Porter to ask what the connection might be. To her
The light surrounding you this very second may be used to expose how much money you make, where you live, when you're home, and much more. That's the big takeaway from A 2016 analysis of ambient light sensors by London-based security and privacy consultant and University College London researcher
24 Apr 2016
As part of its Smart Nation programme, in 2016 Singapore launched the most extensive collection of data on everyday living ever attempted in a city. The programme involved deploying myriad sensors and cameras across the city-state to comprehensively monitor people, places, and things, including all
30 Nov 2015
A 2015 study carried out in Rwanda and published in the journal Science used mobile phone records to study the distribution of wealth and poverty in an attempt to fill in the gap left by the difficulty of collecting accurate statistics. A rought idea of geographic location was derived from
09 Oct 2015
In 2015, Ant Financial, a Chinese company affiliated with Alibaba, began deploying "Sesame Credit scores". Ant, which also runs the popular payment app Alipay, claims that it uses both online and offline purchasing and spending habits to calculate a credit worthiness score for each of its more than
24 Nov 2016
In 2016 researchers in China claimed an experimental algorithm could correctly identify criminals based on images of their faces 89% of the time. The research involved training an algorithm on 90% of a dataset of 1,856 photos of Chinese males between 18 and 55 with no facial hair or markings. Among
04 Jan 2016
In 2016, the US's third-largest property and casualty insurer, Liberty Mutual, announced it would partner with Subaru to enable drivers who have bought Subaru's Starlink infotainment system to download a car app that will notify them if they are accelerating too aggressively or braking too hard. The
A new generation of technology has given local law enforcement officers in some parts of the US unprecedented power to peer into the lives of citizens. The police department of Frenso California uses a cutting-edge Real Time Crime Center that relies on software like Beware. As officers respond to
23 May 2016
Computer programs that perform risk assessments of crime suspects are increasingly common in American courtrooms, and are used at every stage of the criminal justice systems to determine who may be set free or granted parole, and the size of the bond they must pay. By 2016, the results of these
10 Aug 2015
In May 2015, the US Department of Justice and the FBI submitted a declaration to an Oregon federal judge stating that the US government's no-fly lists and broader watchlisting system relied on predictive judgements of individuals rather than records of actual offences. The documents were filed as