\n
![](\"https:\/\/i0.wp.com\/www.rsoftsolution.com\/blog\/wp-content\/plugins\/RSSPoster_PRO\/cache\/a8b04_screen-shot-2016-11-02-at-2-32-33-pm.png?w=640\")
<\/p>\n
\n<\/p>\n
UK insurance firm\u00a0Admiral had intended to launch an app this week offering discounted car insurance premiums to first time drivers based on an algorithmic assessment of their Facebook posts.<\/p>\n
All drivers would have had to do is sign in with their Facebook login\u00a0to\u00a0grant permission to the company to scan their Facebook posts in order to get a potential discount on their car insurance premiums.<\/p>\n
However the experiment\u00a0has fallen foul of Facebook\u2019s platform policy, which puts strict limits on how developers on the\u00a0platform can use the information users\u00a0share with them.<\/p>\n
Clause 3.15 of the\u00a0policy also specifically prohibits\u00a0use of data obtained from Facebook to<\/p>\n
\n\u2026make decisions about eligibility, including whether to approve or reject an application or how much interest to charge on a loan.<\/span><\/p>\n<\/blockquote>\n
In an interview with The Guardian<\/a>\u00a0about the opt-in firstcarquote app, project lead Dan Mines described it\u00a0as \u201ca test\u201d, saying:\u00a0\u201cWe are doing our best to build a product that allows young people to identify themselves as safe drivers\u2026 This is innovative, it is the first time anyone has done this.\u201d<\/p>\n
The algorithms that Admiral\u00a0had developed for the app apparently aimed to\u00a0glean\u00a0personality traits\u00a0from\u00a0users\u2019 Facebook posts by analyzing how posts were written \u2014 with individuals who\u00a0scored well for\u00a0qualities such as conscientiousness and organization more likely to be offered discounts\u00a0vs those who came\u00a0across as overconfident\/less well organized, as judged by their Facebook postings.<\/p>\n
Photos were not intended to be\u00a0used to assess drivers \u2014 the analysis was purely based on text updates to Facebook.<\/p>\n
\u201cOur analysis is not based on any one specific model, but rather on thousands of different combinations of likes, words and phrases and is constantly changing with new evidence that we obtain from the data,\u201d Yossi Borenstein, the principal data scientist on the project, told the paper. \u201cAs such our calculations reflect how drivers generally behave on social media, and how predictive that is, as opposed to fixed assumptions about what a safe driver may look like.\u201d<\/p>\n
Giving a more specific\u00a0example of how Admiral\u2019s app\u00a0would be\u00a0assessing a Facebook user\u2019s attitude\u00a0behind the wheel, The Guardian suggested overuse of exclamation marks in Facebook posts might count against a first time driver, while\u00a0posting lists and writing in short, concrete sentences containing\u00a0specific detail would be seen as a plus.<\/p>\n
Admiral said no price rises would be incurred as a result of using the app but discounts of up to \u00a3350 were set to be offered\u00a0\u2014 although the company was also not ruling out expanding the project\u00a0in future to loop in additional social media services and, potentially, to also increase premiums for some drivers.<\/p>\n
\u201cThe future is unknown,\u201d said Mines. \u201cWe don\u2019t know if people are prepared to share their data. If we find people aren\u2019t sharing their data, then we won\u2019t ever get to consider that [expanding firstcarquote].\u201d<\/p>\n
As it turns out, the app\u2019s future is unknown as Facebook is not prepared to share user data with Admiral for this eligibility assessment\u00a0use-case. Which, if the team had read\u00a0Facebook\u2019s platform policy, should have been immediately clear.<\/p>\n
Presumably\u00a0Admiral has been\u00a0working on the app for multiple\u00a0months at the very least. Yet\u00a0again, any Facebook platform developer should be aware that all apps are subject to\u00a0final review by the company before they can go\u00a0live to ensure compliance with its\u00a0platform\u00a0policy. Even \u201ctest\u201d apps.<\/p>\n
Admiral now says the firstcarquote\u00a0launch has been delayed \u2014 noting on the website <\/a>that: \u201cWe were really hoping to have our sparkling new product ready for you, but there\u2019s a hitch: we still have to sort a few final details.\u201d<\/p>\n
It also touts other use cases for the app \u2014 such as being able to see what some other new drivers have paid for\u00a0car insurance and some\u00a0details of the cars they drive. Although that\u2019s a far cry from offering first time drivers discounts based on how many exclamations marks they typically deploy\u00a0in their Facebook posts.<\/p>\n
We tried to contact the company with questions but at\u00a0the time of writing Admiral had not responded, and its press office had professed itself too busy to speak \u2014 with\u00a0an outside PR firm being engaged to fence queries. We\u2019ll update this story with any response.<\/p>\n
In a statement provided to TechCrunch a Facebook spokesperson confirmed\u00a0Admiral will only be able to use Facebook accounts for login and identity verification \u2014 so not for scanning post data. The company\u00a0further suggests the insurer\u00a0intends to rework the app to create\u00a0an alternative data source to\u00a0assess drivers\u2019 eligibility.<\/p>\n
The Facebook spokesperson said:<\/p>\n
\nWe have clear guidelines that prevent information being obtained from Facebook from being used to make decisions about eligibility.<\/p>\n
We have made sure anyone using this app is protected by our guidelines and that no Facebook user data is used to assess their eligibility. Facebook accounts will only be used for login and verification purposes.<\/p>\n
Our understanding is that Admiral will then ask users who sign up to answer questions which will be used to assess their eligibility.<\/p>\n<\/blockquote>\n
It\u2019s worth\u00a0noting that Facebook has itself patented using social graph for assessing eligibility of creditworthiness, as the Atlantic<\/a> reported last year.<\/p>\n
US patent 9,100,400<\/a>, granted to Facebook in August 2015, includes a specific method for authenticating an individual for access to information or service \u201cbased on that individual\u2019s social network\u201d \u2014 with one\u00a0of the examples given using the scenario of a\u00a0service provider being\u00a0a lender who assesses an individual\u2019s creditworthiness based on the average credit rating of the people the individual is connected to on their social network\u2026<\/p>\n
\nIn a fourth embodiment of the invention, the service provider is a lender. When an individual applies for a loan, the lender examines the credit ratings of members of the individual\u2019s social network who are connected to the individual through authorized nodes. If the average credit rating of these members is at least a minimum credit score, the lender continues to process the loan application. Otherwise, the loan application is rejected.<\/p>\n<\/blockquote>\n
It\u2019s unclear whether Facebook intends or intended to launch any such creditworthiness assessment service itself \u2014 we asked and it did not respond. But many patents are filed defensively and\/or speculatively. And,\u00a0as the Atlantic notes,\u00a0using a person\u2019s social graph\u00a0to assess creditworthiness would\u00a0run huge risks of\u00a0attracting discrimination lawsuits. So the patent does not really read like a serious product proposal on Facebook\u2019s part.<\/p>\n
Beyond that, if Facebook\u2019s platform were to become implicated in weighty\u00a0external\u00a0assessments of individuals, with the potential to have seriously\u00a0negative impacts on their lives, the company would risk discouraging users from\u00a0sharing the sort of\u00a0personal\u00a0data its ad-targeting\u00a0business model relies on. Which is surely part of the reason it\u2019s denying\u00a0Admiral\u00a0the ability to scan\u00a0Facebook posts to assess driving proficiency.<\/p>\n
Facebook is already negatively implicated in state surveillance activity as a\u00a0honeypot\u00a0of data\u00a0utilized by\u00a0intelligence and law enforcement agencies. And on privacy grounds, given its own business model relies on profiling users for ad targeting. But stepping into offering formal assessments of individuals\u2019 creditworthiness, for example, would feel like\u00a0a massive pivot for the social media giant \u2014 although the temptation for it to try to unlock more \u2018worth\u2019 from\u00a0the mountain of data it sits on is\u00a0only set\u00a0to grow, given\u00a0AI\u2019s\u00a0rising star and growing appetite for data.<\/p>\n
In a blog post welcoming Facebook blocking Admiral from scanning users\u2019 posts, digital rights organization the Open Rights Group<\/a>\u00a0points\u00a0out the underlying biases that can\u00a0make any such algorithmic assessments problematic.<\/p>\n
\u201cThere are significant risks in allowing the financial or insurance industry to base assessments on our social media activity,\u201d writes Jim Killock. \u201cWe might be penalised for our posts or denied benefits and discounts because we don\u2019t share enough or have interests that mark us out as different and somehow unreliable. \u00a0Whether intentional or not, algorithms could perpetuate social biases that are based on race, gender, religion or sexuality. Without knowing the criteria for such decisions, how can we appeal against them?\u201d<\/p>\n
\u201cInsurers and financial companies who are beginning to use social media data need engage in a public discussion about the ethics of these practices, which allow a very intense examination of factors that are entirely non-financial,\u201d he adds.<\/p>\n
\nAsked for his\u00a0view on the risks of Facebook itself using its platform to sell assessments on the fitness of its users for accessing other products\/services, such as financial products, Killock also told TechCrunch: \u201cRules on profiling and use of data have to ensure that people are not disadvantaged, unfairly judged, or discriminated against. Facebook\u2019s data is rich, but often ambiguous, may lack context and presents many risks. It is not clear to us that social media information is an appropriate tool for financial decision making.\u201d<\/p>\n
Also blogging<\/a>\u00a0about\u00a0Admiral\u2019s attempt to turn Facebook data into premium-affecting\u00a0personality assessments, law professor Paul Bernal\u00a0voices similar concerns about what he dubs the\u00a0\u201cvery significant\u201d\u00a0risks of such a system being discriminatory.<\/p>\n
\u201cAlgorithmic analysis, despite the best intentions of those creating the algorithms, are not neutral, but embed the biases and prejudices of those creating and using them,\u201d he writes. \u201cA very graphic example of this was unearthed recently, when the first international beauty contest judged by algorithms managed to produce remarkably prejudiced results<\/a> \u2013 almost all of the winners were white, despite there being no conscious mention of skin colour in the algorithms.\u201d<\/p>\n
Bernal\u00a0also argues that the sort of linguistic analysis Admiral\u2019s app was apparently intending would have\u00a0\u201cvery likely\u201d \u00a0favored Facebook users in command of \u201cwhat might be seen as \u2018educated\u2019 language \u2013 and make any kind of regional, ethnic or otherwise \u2018non-standard\u2019 use of language put its user at a disadvantage\u201d.<\/p>\n
\u201cThe biases concerned could be racial, ethnic, cultural, regional, sexual, sexual-orientation, class-based \u2013 but they will be present, and they will almost certainly be unfair,\u201d he adds.<\/p>\n
Bernal goes on to suggest that\u00a0Facebook\u00a0users develop \u201csurvival tactics\u201d as a short term fix for\u00a0defeating\u00a0any assessments being made of them based on their social graphs and footprints \u2014 urging especially young\u00a0people\u00a0(who are perhaps currently most at risk of being harmfully judged by their social media activity) to\u00a0\u201ckeep the more creative sides of your social life off Facebook\u201d.<\/p>\n
He also calls for\u00a0a push by regulators towards developing a framework for\u00a0algorithmic accountability to control the autonomous technologies being increasingly deployed\u00a0to control us.<\/p>\n
\u201cAlgorithms need to be monitored and tested, their impact assessed, and those who create and use them to be held accountable for that impact,\u201d he adds. \u201cInsurance is just one example \u2013 but it is a pertinent one, where the impact is obvious. We need to be very careful here, and not walk blindly into something that has distinct problems.\u201d<\/p>\n
Algorithmic accountability was also\u00a0flagged as a concern by\u00a0a UK science and technology parliamentary committee\u00a0last month<\/a>, in a report considering the \u201chost of social, ethical and legal questions\u201d that arise from growing use of autonomous technologies, and given how quickly machine learning algorithms are being deployed to\u00a0wrangle\u00a0insights from data-sets.<\/p>\n
The committee recommended the government establishes a standing Commission on Artificial Intelligence aimed at \u201cidentifying principles to govern the development and application of AI\u201d, and to provide advice and encourage public dialogue about automation\u00a0technologies.<\/p>\n