Category: blogpost

Saving the Future of Phone Calls – The Fight to Stop Robocalls
By Anonymous | July 5, 2019

“Hello, this is the IRS. I am calling to inform you of an urgent lawsuit! You are being sued for failing to pay taxes and we have a warrant out for your arrest. Please call this number back immediately!”

The familiar noisy background laced with thinly veiled threats is a message many are unfortunately accustomed to. Robocalls are a pervasive annoyance that has become the top consumer complaint to the Federal Trade Commission (FTC). And despite robocalls being prohibited by law, Americans were bombarded by a record breaking 4.4 billion robocalls in June 2019. That’s 145 million calls per day, 13 calls per person!

Figure 1: YouMail Robocall Index: https://robocallindex.com/

So, how does robocallers obtain phone number anyways? Most often, they acquire numbers from third party data providers, who in turn acquired numbers from a variety of avenues that everyday users may not be aware are collecting and selling their data. Some of these sources include:

• Toll free (1-800) numbers that employ caller ID which can collect phone numbers
• Entries into contests where users provided phone numbers in the process
• Applications for credit
• Contributions to charities where users provided phone numbers in the process

Methods of manipulating users into giving up personal information have evolved over the years as well. Robocalls can disguise their numbers to appear as a local telephone number with neighboring area codes to trick users into picking up unfamiliar calls outside of their personal contacts. The variety of robocallers disguising themselves as government agencies, municipal utility providers, or even hospital staff to scam users into providing personal information has grown to such an astonishing extent that lawmakers are now paying attention.

Figure 2: FTC Phone Scams: https://www.consumer.ftc.gov/articles/0076-phone-scams

In November 2018, the Federal Communications Commission (FCC) called on carriers to develop an industry-wide standard to screen and block robocalls. In particular, the FCC urged carriers to adopt the SHAKEN (Secure Handling of Asserted information using toKENs) and STIR (Secure Telephone Identity Revisited) frameworks by the end of 2019. In particular, SHAKEN/STIR frameworks employs secure digital certificates to validate that calls are from the purported source and has not been spoofed. Each telephone service provider must obtain a digital certificate from a certified authority and this enables called parties to verify the accuracy of the calling number.

Furthermore, in January 2019, Senators Edward J. Markey and John Thune introduced the Traced Act that aims to require all telephone service providers, including those over the internet such as Google Voice or Skype, to adopt similar call authentication technologies.

Together, the collective drive by private industry and regulatory efforts will make it harder for the majority of robocallers to spam consumers at the touch of a button. Like spam emails, calls with suspicious or unverified origins can be traced and blocked en masse. And though these recent tactics are certainly a step in the right direction for consumer protection, some fear that historically underserved communicated might not upgrade in time and be risk being further isolated. Rural areas that often rely on older landlines will foreseeably struggle to adopt the new technology due to outdated equipment and cost to implement. Immigrant communities who make and receive international calls to foreign countries might be subjected to higher levels of discrimination as international calls cannot yet be fully authenticated. This means their calls may be more likely to be labeled as fraud and increased targeting by robocall operatives that will exploit this gap in technology to scam an already vulnerable population.

As the world continues to evolve with newer technology, it’s important to not only think about who will benefit from these changes, but also who will be left behind. In this case, as the FCC and private industry work together to protect consumers, they should also seek to mitigate the risk of scam and spam robocalls targeting vulnerable communities. One way to accomplish this is to work with other regulatory agencies, such as the Housing and Urban Development department, to create long term and sustainable incentives within rural areas to modernize their infrastructure. Another way is for private industries who are vested in international businesses to continue working closely with regulators to develop a global SHAKEN/STIR standard that protects an increasingly globalized world. Afterall, robocalls are hardly a uniquely American phenomenon. However, taking the lead in safeguarding the next generation can be a defining American trademark.

 

Bibliography

• “How Do Robo-Callers and Telemarketers Have My Cell Number Anyway?” BBB, www.bbb.org/acadiana/news-events/news-releases/2017/04/how-do-robo-callers-and-telemarketers-have-my-cell-number-anyway/.
• “How to Know It’s Really the IRS Calling or Knocking on Your Door.” Internal Revenue Service, www.irs.gov/newsroom/how-to-know-its-really-the-irs-calling-or-knocking-on-your-door
• “Phone Scams.” Consumer Information, 3 May 2019, www.consumer.ftc.gov/articles/0076-phone-scams.
• “Thune, Markey Reintroduce Bill to Crack Down on Illegal Robocall Scams.” Senator Ed Markey, 17 Jan. 2019, www.markey.senate.gov/news/press-releases/thune-markey-reintroduce-bill-to-crack-down-on-illegal-robocall-scams.
• Vigdor, Neil. “Want the Robocalls to Stop? Congress Does, Too.” The New York Times, The New York Times, 20 June 2019, www.nytimes.com/2019/06/20/us/politics/stopping-robocalls.html.
• “YouMail Robocall Index: June 2019 Nationwide Robocall Data.” Robocall Index, robocallindex.com/.

Audit organizations, trust and their relationship with ethical automated decision making
By Jay Venkata | July 5, 2019

The world runs on trust. Worldwide billions of dollars are spent every year on developing, and maintaining trust. Any transaction, whether it be supply chain, finance or healthcare related requires trust between people, businesses and entities. As an individual consumer, you are making decisions based on trust on an almost hourly basis. This goes from trusting the safety of your meals to trusting the financial transactions done through your bank. Audit organizations and regulators, both private and public, are responsible for maintaining this trust in society. I work at one of the Big 4 global audit firms. At the core of what each of these audit companies do is giving assurance to businesses and governments. The mission statement of my company actually is ‘Solving complex problems and building trust in society’. But what does trust look like in this digital world?

[Image 1]

Trust in the Digital World

In yesteryears, audit organizations would primarily base their decisions on financial ledgers, and sources of decisions can be razored down to select executives or managers. Manual and paper based processes could only be kept track of manually. However the trend towards automating business processes, and their associated accounting and strategic decisions is causing an interesting challenge for regulators. A realm of work historically led by humans such as deciding on credit card applications can increasingly be automated. There is now a need for alternative methods to develop the same level of trust again. One solution to this issue is to focus more on independently auditing the underlying algorithms. Audit firms may need to have technical staff who can work alongside the functional experts to decode the algorithms and get to the root of any errors or biases that could affect the decisions and outcomes. Hence there is a need for accounting colleges across the world to focus on these interdisciplinary skills that will make students more ready for their careers post graduation. Another challenge is that most businesses and governments do not seem very willing to publish algorithms, the data used to train them or the inferences made from the data.

[Image 2]

Auditing the algorithms

A longer term solution that could be effective is to work alongside the government to create transparency and openness standards that are applicable to all organizations. These types of guardrails exist already in financial statements and reporting, which are managed closely by the SEC in the US. The GDPR currently has a requirement to use “appropriate mathematical or statistical procedures” to avoid or reduce risk resulting from errors or inaccuracies. The French administration also announced that algorithms developed for the government use will be made publicly available, so that society at large can verify their correct application. There needs to be a similar push worldwide for rigorous standards on automated processes and decision making to create algorithmic accountability.

[Image 3]

Blockchain improves trust in transactions through distributed ledgers

This trend towards improving and automating trust could happen naturally as we move towards technologies like Internet of Things and Blockchain, which will create end-to-end traceability for products and transactions in a cheap and ubiquitous manner. However the case for auditing algorithms is clear. Audit firms and regulators need to be one step ahead of the organizations they are auditing at all times and this applies to the current scenario where the stakes couldn’t be higher- ensuring the integrity of data flow and decision making.

Works Cited

• Abraham, C., Sims, R. R., Daultrey, S., Buff, A., & Fealey, A. (2019, March 18). How Digital Trust Drives Culture Change. Retrieved July 7, 2019, from https://sloanreview.mit.edu/article/how-digital-trust-drives-culture-change/
• O’Neil, C., & Schermer, B. (2018, July 30). Audit the algorithms that are ruling our lives. Retrieved July 7, 2019, from https://www.ft.com/content/879d96d6-93db-11e8-95f8-8640db9060a7
  What is Blockchain Technology? (2018, September 11). Retrieved July 7, 2019, from https://www.cbinsights.com/research/what-is-blockchain-technology/
• Likens, S., & Bramson-Boudreau, E. (2019, May 02). Blockchain Promises Trust, But Can It Deliver? Presented by PwC – MIT Technology Review. Retrieved July 7, 2019, from https://events.technologyreview.com/video/watch/pwc-blockchain-trust-likens/

The Role of Providers in Free Flowing data
By Hanna Rocks | July 5, 2019

In the past year, we have all experienced an inundation of emails with the subject line, “We’ve updated our privacy policies”. Recent actions by the Federal Trade Commission [1] (FTC) and the European Union (does the acronym ‘GDPR’ [2] ring any bells?) have prompted companies to make significant changes to the way they manage user information.

Despite all these changes to policies and headlines about what they hold, the majority of users continue to scroll to the bottom, check “I accept”, and move on without taking any time to consider what they are agreeing to. In fact, a study published in 2016 and updated in 2018 [3] found that 97% of users agreed to the privacy policy of a fictitious social media site—glossing over the clause requiring provision of the user’s first-born child as payment for the service.

Why are we so willing to accept any terms to gain access to an online service? The answer likely lies within the value that users receive in exchange for clicking “I accept”. A Deloitte survey [4] of over 8,500 consumers across six countries found that 79% were willing to share their personal data, *so long as there was a clear benefit to the user.*

This stance aligns with many legal and ethical opinions on privacy protections. One of the earliest regulatory frameworks on privacy harms, the Belmont Report [5], clearly states that the organization should weigh the potential for harm against the possible benefits to the individual. However, because the perceived benefit will vary from user to user, this value is difficult to define or estimate.

Instagram, for example, may define “benefit” as providing specific, personalized ad content to each user. This has also led to a heated debate about whether or not Instagram (or any other app on your phone) is “listening” to us [6]. I often wonder about this, but my worries have weakened over the years as I consume of a growing number of products found while scrolling through my feed. The value I receive from these products is worth whatever information Instagram has been collected and analyzed. I don’t know the details, so it is easy not to care.

Which brings us back to regulatory bodies like the FTC. The FTC is tasked with protecting ignorant or lazy consumers from corporations who are after unreasonable amounts of personal data. However, the question of what is deemed “reasonable” will continue to change as companies differentiate themselves by providing the ultimate personalized customer experience. More and more consumers are coming to expect tailored recommendations from the services they use—whether that is the “perfect” new pair of shoes discovered on Instagram or a carefully calculated rate from your insurance company.

As we continue down the path of highly customized goods and services, it is critical that businesses appoint individuals, or even teams, to provide oversight of what data the organization collects from its consumers and how that data is used. Doing so will benefit both the consumer and the provider by monitoring policies and comparing those policies to existing or emerging regulatory frameworks. Businesses would do well to adopt a proactive approach…

such as the “opt in” requirement under GDPR [7] that clearly addresses how they use customer data, rather than expecting consumers to read thousands of words written in the dreaded “legal-ese”. If a business is willing to invest time and resources, it *can* provide ultimate customization with ultimate protection. In the end, a business that brands itself as a leader in the responsible use of consumer data will surely attract more customers—offering both a personalized experience and peace of mind.

References:
[1]: https://www.vox.com/2019/1/23/18193314/facebook-ftc-fine-investigation-explained-privacy-agreement
[2]: https://www.techrepublic.com/article/the-eu-general-data-protection-regulation-gdpr-the-smart-persons-guide/
[3]: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2757465
[4]: https://www2.deloitte.com/insights/us/en/industry/retail-distribution/sharing-personal-information-consumer-privacy-concerns.html
[5]: https://www.hhs.gov/ohrp/regulations-and-policy/belmont-report/read-the-belmont-report/index.html#xassess
[6]: https://www.vox.com/the-goods/2018/12/28/18158968/facebook-microphone-tapping-recording-instagram-ads
[7]: https://www.cooleygo.com/gdpr-do-i-need-consent-to-process-personal-data/

Search Engines and Reporting Illegal Activity
By Anonymous | July 5, 2019

Search engines are one of the most ubiquitous internet services and generate some of the largest databases in the world as a result. In 2012, Google handled over 1 trillion search queries [1]. However, a certain subset of such search queries point to less benign actions, and may be important clues to help identify terrorists, child abusers, human traffickers, and illegal drug traders. Should search engines be obligated to report on the potential for such illegal activity so it can be followed up on?

First we should examine how easy this would be from a technical standpoint. To do this, I went through the search logs of a major search engine to try and find a query I had input earlier.

Figure 1: Finding a user in search data: surprisingly easy.

Within a few minutes I had my nonsensical query (Column 4), the URLs I clicked (Column 6), my rough location information (Column 11), and my IP address (Column 21) which is considering personally identifying under GDPR rules [2]. If a search company were to target for other queries like “how to make a bomb” they could quickly collect the IP addresses of potential lawbreakers and cross reference with ISP companies to figure out names and home addresses. The ease of success of this experiment indicates the technical costs of implementing a filtering and reporting system to be not be a blocking issue.

There is already some basis for the idea that search engines should be obligated to report on suspected illegal activity especially regarding child abuse. In the US, all states have mandatory reporter laws that require professionals which may be seen as trusted sources for children (such as teachers, health-care workers, and child care providers) to make a report if they suspect that a child has been abused or neglected [3]. A similar law exists for companies to report child sexual abuse, which Google used in 2014 after finding explicit photos in a scan of a user’s Gmail account [4]. The law, importantly, does not require companies to be proactive and monitor or affirmatively search for illegal content which would be easy for companies to implement [5].

Figure 2: Google already searches through some user data and is mandated to report child pornography it finds.

The main drawbacks to search engines entering the space of reporting illegal activity are the concerns over a users rights and expectations to privacy. In Gonzales v. Google Inc, the U.S government subpoenaed Google and other search engine companies for queries and URLs for a study [6]. The court ruled that forcing Google to divulge query and url data would diminish user trust in the company due to the fact that users have some expectation of privacy when using the platform. It is this expectation of privacy that keeps many users from migrating away from major search engines to smaller privacy concious alternatives.

There are also issues with false positives, where users may search terms jokingly or out of curiosity and trigger automated alerts leading to unfounded accusations. This actually happened in 2013 when a counter-terrorism unit mistakenly searched a family’s home after the husband searched “pressure cooker bomb” and “backpacks” out of curiosity at work [7]. If search engines stepped into this space, law enforcement agencies may be faced with an overwhelming number of false positive cases which would waste resources and cause bad press.

Figure 3: Searching for pressure cookers and backpacks could trigger a false positive and get your house searched.

Finally, there is the argument that searches are protected under free speech. In United States v. Gilberto Valle, the US Second Circuit Court of Appeals reversed the ruling that Valle’s Google searches, which were related to fantasizing about violent crimes, themselves constituted a crime. The court reasoned that no action was taken thus no actual crime was committed, and that Valle was free to express his fanatasies [8]. This ruling was supported by a number of First Amendment and Internet law scholars and points to search queries being protected as free speech, making it more of a potential PR disaster for search engine companies to be handing it off to law enforcement or the government.

Privacy expectations and laws surrounding search engines are still a hotly debated and contested area today. The technologies to allow search engines to find and report suspected cases of illegality are easy to implement, and there is some movement in the legal sphere to cover cases like child abuse where it is clear that search engines are ethically obligated to act. On the other side, there are genuine privacy and trust concerns, false positive issues, and legal battles over the coverage of free speech that will probably keep search engine companies sidelined. Most likely, these companies will wait for the legal contests to be settled before deciding to move into any reporting of potential criminal activities unless explicitly forced to by law, so it up to all of us to decide whether the tradeoff is worth it or not.

References

[1] https://archive.google.com/zeitgeist/2012/#the-world
[2] https://eugdprcompliant.com/personal-data/
[3] https://www.childwelfare.gov/pubPDFs/manda.pdf
[4] https://mashable.com/2014/08/04/google-gmail-child-porn-case/
[5] https://www.law.cornell.edu/uscode/text/18/2258A
[6] https://scholarship.law.berkeley.edu/cgi/viewcontent.cgi?article=1689&context=btlj
[7] https://www.theguardian.com/commentisfree/2013/aug/01/government-tracking-google-searches
[8] https://www.eff.org/cases/united-states-v-gilberto-valle

HR Analytics – An ethical dilemma?
By Christoph Jentzsch | June 28, 2019

In the times of “Big Data Analytics”, the “War for Talents” and “Demographic Change” the key words HR Analytics or People Analytics seem to be ubiquitous in the realm of HR departments. Many Human Resource departments ramping up their skills on analytical technologies to deploy the golden nuggets of data they have about their own workforce. But what does HR Analytics even mean?

Mick Collins, Global Vice President, Workforce Analytics & Planning Solution Strategy at SAP SuccessFactors sets the context of HR Analytics as the following:

“The role of HR – through the management of an organization’s human capital assets – is to impact four principal outcomes: (a) generating revenue, (b) minimizing expenses, (c) mitigating risks, and (d) executing strategic plans. HR analytics is a methodology for creating insights on how investments in human capital assets contribute to the success of those four outcomes. This is done by applying statistical methods to integrated HR, talent management, financial, and operational data.” (Lalwani, 2019)

Summarized, HR Analytics is a data-driven approach to HR Management.

Figure 1: Data-Driven Decision Making in HR – Source: https://www.analyticsinhr.com/blog/what-is-hr-analytics/

So, what’s the big fuzz about it? Well, the example of Marketing Analytics, which had a revolutionary impact on the field of marketing, showcases that HR analytics is changing the way how HR departments operate tomorrow. A more data-driven approach enables HR to:

• …make better decisions using data, instead of relying on the managers gut feeling
• …move from an operational partner to a tactical, or even strategic partner (Vulpen, 2019)
• …attract more talent, by improving the hiring processes and the employee experience
• …continuously improve workforce planning through informed talent development (MicroStrategy Incorporated, 2019)

However, the increased availability of new findings and information as well as the ongoing digitalization that unlocks new opportunities to understand and interpret those information, also raises new concerns. The most critical challenges are:

• Having employees in HR functions with the right skillset to gather, manage, and report on the data
• Confidence in data quality as well as cleansing and interpretation problems
• Data privacy and compliance risks
• Ethical and moral concerns about using the data

Especially, the latter 2 aspects are up for investigation and some guidance is given on how to overcome those challenges. It is important to understand in the first place that corporate organizations collect data about their employees on very detailed level. Theoretically they could reconcile findings back down the level that identifies an individual employee. However, in the considerations of legal requirements this is not always allowed and due to the implementation of GDPR regulations, organizations are now forced to look at employee data and privacy in the same way they do for customers.

Secondly, it is crucial to understand that HR Analytics uses a range of techniques based on statistics that are incredibly valuable at the population level but they can be problematic if you use them to make a decision about an individual. (Croswell, 2019)

Figure 2: HRForecast Recruiting Analytics Dashboard source: https://www.hrforecast.de/portfolio-item/smartinsights/

This is being confirmed by Florian Fleischmann, CEO of HR Analytics Provider HRForecast as he states: “The real lever of HR Analytics is not taking place on an individual employee level, it is instead happing on a corporate macro level, when organizational processes, such as the hiring procedure or overarching talent programs are being improved.” (Fleischmann, 2019). Mr. Fleischmann is totally right, as managing people on an individual level is still a person-to-person relationship between employee and manger, which is nothing that requires a Big Data algorithm. Assuming the worst-case scenario: Job Cuts. If low-performers ought to be identified, simply line managers have to be interviewed – there is no need for a Big Data solution.

Analytics on an individual level do not bring added value but can even create harm as Mr. Fleischmann points out: “According to our experience the application of AI technology to predict for example, employee attrition rates on an individual basis can create more harm than benefit. It can cause a self-fulfilling prophecy, as the manager believes to know what team member is subject to leave and changes his behavior accordingly in a negative way”. (Fleischmann, 2019)

For that reason, HRForecast advocates for two paradigms in the ethical use and application for HR Analytics:

1. Information on an employee level is only provided to the individual employee and is not shared with anyone else. “This empowers the employee to stay performant as he or she can analyze for example his or her own skill set against a benchmark of skills that are required in the future”, confirms Fleischmann.
2. Information is being shared with management only on an aggregated level. The concept of “Derived Privacy” is applicable in this context as it allows enough insights to draw conclusions on a bigger scale but protects the individual employee. Given the legal regulations data on that level needs to be fully anonymized and groups smaller than 5 employees are excluded from any analysis. Fleischmann adds: “The implementation of GDPR did not affect HRForecast, as we applied those standards already pre-GDPR. Our company stands to a high ethical code of conduct, which is a key element if you want to be a successful player in the field of HR Analytics.”

In conclusion it can be stated that the application of Big Data Analytics or AI in a context of Human resources can create a huge leap in organizational transparency. However, his newly won information can cause major privacy risks for employees if not treated in reasonable fashion. To mitigate the risk of abusing the increased level of transparency an ethical code of conduct as provided by a third-party expert HRForecast needs to be applied in modern organizations. Thus, Big Data in HR can lead to an ethical dilemma, but it does not have to.

Bibliography

• Croswell, A. (2019, June 25). Why we must rethink ethics in HR analytics. Retrieved from Why we must rethink ethics in HR analytics: https://www.cultureamp.com/blog/david-green-is-right-we-must-rethink-ethics-in-hr
• Fleischmann, F. (2019, June 25). CEO HRForecast. (C. Jentzsch, Interviewer)
• Lalwani, P. (2019, April 29). What Is HR Analytics? Definition, Importance, Key Metrics, Data Requirements, and Implementation. Retrieved from What Is HR Analytics? Definition, Importance, Key Metrics, Data Requirements, and Implementation: https://www.hrtechnologist.com/articles/hr-analytics/what-is-hr-analytics/
• MicroStrategy Incorporated. (2019). HR Analytics – Everything You Need to Know. Retrieved from HR Analytics – Everything You Need to Know: https://www.microstrategy.com/us/resources/introductory-guides/hr-analytics-everything-you-need-to-know
• Vulpen, E. v. (2019). HR Analytics. Retrieved from What is HR Analytics?: https://www.analyticsinhr.com/blog/what-is-hr-analytics/

Privacy Movements within the Tech Industry
By Jill Rosok | June 24, 2019

An increasing number of people have become fed up with major tech companies and are choosing to divest from companies that violate their ethical standards. There’s been #deleteuber, #deletefacebook and other similar boycotts of big tech companies that violated consumer trust.

In particular, five companies have an outsized influence on the technology industry and the economy in general, Amazon, Apple, Facebook, Google (now a unit of parent company, Alphabet) and Microsoft. Among numerous scandals, Facebook has insufficiently protected user data leading to Russian hacking and the Cambridge Analytica controversy. Amazon and Apple have been chastised for unsafe working conditions in their factories. Google contracts with the military and collects a massive amount of data on their users. Microsoft has been repeatedly involved in antitrust suits. Those who have attempted to eliminate the big five tech companies from their lives have found it nearly impossible. It’s one thing to delete your Facebook and Instagram accounts and stop ordering packages from Amazon. However, eliminating the big five companies from your life is actually much more complicated than that. The vast majority of smartphones have hardware and/or software built by Apple and Google. Furthermore, Amazon’s services run the backend of a huge number of websites, meaning stepping away from these companies would essentially mean to stop using the internet.

For a limited few, it might be possible to simply log off and never come back, but most people rely on tech companies in some capacity to provide them basic access to work, connection to friends and family, and the internet in general. As the big five acquire more and more services that encompass the entirety of people’s lives it is extremely difficult for an individual to participate in a meaningful boycott of all five companies.

In light of the dominance of these five companies, to what extent is the government responsible for some kind of intervention? And if the government were to intervene, what might this look like? Antitrust legislation is intended to protect the consumer from monopoly powers. Historically, the government’s focus has been ensuring that companies are not behaving in such a way that leads consumers to pay higher prices for goods and services. However, this doesn’t protect users where no cash is exchanged, as in the case of Facebook. It’s a great example of the classic adage, if you’re not paying money for a service, then that means the product is you. It also does not hold up in a circumstance where venture backing or other product lines in the business can enable companies to artificially deflate prices below cost for many years until all other competitors are wiped off the map. Senator and presidential candidate, Elizabeth Warren, recently proposed breaking up big tech. While her piece was received more symbolically than as a fully formed plan to regulate the tech industry, there were aspects that appear to have resonated strongly with the general public. In particular, the idea that mergers and acquisitions by large companies should undergo much deeper scrutiny and perhaps be banned entirely was well received by analysts.

Like with most complex problems in life, there are no easy solutions to simultaneously protect consumers and maximize technological innovation. However, it is vital to avoid becoming stagnant in response to the scale of the problem. Rather, as individuals, we must remain informed and put pressure on our political leaders to enact meaningful legislation to ensure the tech industry does not violate the basic rights of consumers.

Breast Cancer, Genetic Testing, and Privacy
By Anna Jacobson | June 24, 2019

5%-10% of breast cancer is believed to be hereditary, meaning that it results directly from a genetic mutation passed on from a parent. The most common known cause of hereditary breast cancer is an inherited mutation in the BRCA1 or BRCA2 gene; about 70% of women with these mutations will develop breast cancer before the age of 80. Identification of these mutations can determine a breast cancer patient’s course of treatment and post-treatment monitoring, inform decisions about if and how she has children, and raise awareness in her family members of their potentially higher risk.

Because of this, newly diagnosed breast cancer patients may be referred for genetic risk evaluation if they meet criteria laid out in the National Comprehensive Cancer Network (NCCN) genetic testing guidelines, including family medical history, tumor pathology, ethnicity, and age. These at-risk patients typically undergo multi-gene panel testing that looks for BRCA1 and BRCA2 mutations, as well as a handful of other less common gene mutations, some of which are associated with inherited risk for other forms of cancer as well as breast cancer.

Genetic testing for breast cancer is a complex issue that raises many concerns. One concern is that not enough patients have access to the testing; some recent studies have shown that the genetic testing guideline’s criteria are too restrictive, excluding many patients who in fact do carry hereditary gene mutations. Another concern is that the testing is not well-understood; for example, patients and even doctors may not be aware that there are many BRCA mutations that are not detected by current tests, including ones that are more common that those that are currently tested. Yet another set of concerns revolves around the value of predictive genetic testing of family members who do not have a positive cancer diagnosis, and whether the benefit of the knowledge of possible risk outweighs the potential harms.

To help a patient navigate this complexity, this genetic testing is ideally offered in the context of professional genetic expertise for pre- and post-test counseling. However, under a 2013 Supreme Court ruling which declared that genes are not patentable, companies like 23andMe now offer direct-to-consumer BRCA testing without professional medical involvement or oversight. And even at its best, genetic counseling comes at a time at which breast cancer patients and their caregivers may be least able to comprehend it. They may be suffering from the shock of their recent diagnoses. They may be overwhelmed by the vast amount of information that comes with a newly diagnosed illness. Most of all, they may only be able to focus on the immediate and urgent need to take the steps required to treat their disease. To many, it is impossible to think about anything other than whether the test results are positive, and if they are, what to do.

But to a breast cancer survivor, other concerns about her genetic testing may arise months or years later. One such concern may be about privacy. Genetic testing for breast cancer is not anonymous; as with all medical testing, the patient’s name is on the test order and the results, which then become part of the patient’s medical record. All medical records, including genetic test results, are protected under HIPAA (Health Insurance Portability and Accountability Act of 1996). However, the recent proliferation of health data breaches from cyberattacks and ransomware has given rise to growing awareness that the confidentiality of medical records can be compromised. This in turn leads to the fears that exposure of a positive genetic test result — one that suggests increased lifetime cancer risk — could lead to discrimination by employers, insurers, and others.

In the United States, citizens are protected against such discrimination by GINA (Genetic Information Nondiscrimination Act of 2008), which forbids most employers and health insurers from making decisions based on genetic information. However, GINA does not apply to small businesses (with fewer than 15 employees), federal and military health insurance, and other types of insurance, such as life, disability, and long-term care. It also does not address other settings of potential discrimination, such as in housing, social services, education, financial services and lending, elections, and legal disputes. Furthermore, in practice it could be very difficult to prove that discrimination prohibited by GINA took place, particularly in the context of hiring, in which it is not required that an employer give complete or truthful reasons – or sometimes, any reasons at all – to a prospective employee for why they were not hired. And perhaps the greatest weakness of GINA, from the standpoint of a breast cancer survivor, is that it only prohibits discrimination based on genetic information about someone who has not yet been diagnosed with a disease.

Though not protected by GINA, cancer survivors are protected by the Americans with Disabilities Act (ADA), which prohibits discrimination in employment, public services, accommodations, and communications based on a disability. In 1995, the Equal Employment Opportunity Commission (EEOC) issued an interpretation that discrimination based on genetic information relating to illness, disease, or other disorders is prohibited by the ADA. In 2000, the EEOC Commissioner testified before the Senate that the ADA “can be interpreted to prohibit employment discrimination based on genetic information.” However, these EEOC opinions are not legally binding, and whether the ADA protects against genetic discrimination in the workplace has never been tested in court.

Well beyond existing legislative and legal frameworks, genetic data may have implications in the future of which we have no conception today, more than perhaps any other health data. The field of genomics is rapidly evolving; it is possible that a genetic mutation that is currently tested because it signals an increased risk for ovarian cancer might in the future be shown to signal something completely different and possibly more sensitive. And unlike many medical tests which are relevant at the time of the test but have decreasing relevance over time, genetic test results are eternal, as true on the day of birth as on the day of death. Moreover, an individual’s genetic test results can provide information about their entire family, including family members who never consented to the testing and family members who did not even exist at the time the test was done.

The promise of genetic testing is that it will become a powerful tool for doctors to use in the future for so-called “precision prevention”, as well as personalized, targeted treatment. However, in our eagerness to prevent and cure cancer, we must remember to consider that as the area of our knowledge grows, so too grows its vulnerable perimeter – and so must our defenses against those who might wish to misuse it.

________________
References

• “Genetic Testing and Privacy.” Breastcancer.org, 28 Sept. 2016, www.breastcancer.org/symptoms/testing/genetic/privacy.
• “Genetic Testing Guidelines for Breast Cancer Need Overhaul.” Clinicaloncology.com, 24 August 2018, https://www.clinicaloncology.com/Breast-Cancer/Article/08-18/Genetic-Testing-Guidelines-for-Breast-Cancer-Need-Overhaul/52544?sub=–esid–&enl=true.
• “Genetic Information Privacy.” Eff.org. https://www.eff.org/issues/genetic-information-privacy.
• “Genetic Discrimination.” Genome.gov. https://www.genome.gov/about-genomics/policy-issues/Genetic-Discrimination.
• “NCCN Guidelines Version 3.2109.” NCCN.org. https://www.nccn.org/professionals/physician_gls/pdf/genetics_screening.pdf.
• “Understanding Genetic Testing for Cancer.” Cancer.org. https://www.cancer.org/cancer/cancer-causes/genetics/understanding-genetic-testing-for-cancer.html.

Maintaining Data Integrity in an Enterprise
By Keith Wertsching | June 21, 2019

Everyone suffers when an enterprise does not maintain the integrity of its data and the leaders employ that data to make important decisions for the enterprise. There are many roles involved in mitigating the risk of poor data integrity, which is defined by Digital Guardian as “the accuracy and consistency (validity) of data over its lifecycle.” But who should be responsible for making sure that the integrity of the data is preserved throughout collection, extraction, and use by the data consumers?
The agent who maintains data accuracy should ideally be someone who:

• Understands where the data is collected from and how it is collected
• Understands where and how the data is stored
• Understands who is accessing the data and how they are accessing it
• Has the ability to recognize when that data is not accurate and understands the steps required to correct it

Too often, the person responsible for maintaining data integrity is focused primarily on the second bullet point, with a casual understanding of the first and third bullet points. Take this job description for a data integrity analyst from Investopedia:
“The primary responsibility of a data integrity analyst is to manage a company’s computer data by way of monitoring its security…the data integrity analyst tracks records indicating who is accessing what information held by company computer systems at specific times.”

The job description demonstrates that someone working in data integrity should be an expert on where and how the data is stored, and be familiar with who should be accessing that information in order to make sure that company data is not stolen or used inappropriately. But who is ultimately responsible for making sure that the information is accurate in the first place, and for making sure that any changes needed are done in a timely fashion and tracked for future records?

In today’s world of enterprise database administrators, there is often a distinct separation between the person or team that understands how the data is stored and maintained and the person or team that has the ability to recognize when the data is not accurate. Let’s take the example of a configuration management database (CMDB) to highlight the potential issues from separation of data integrity responsibility. SearchDataCenter defines a CMDB as “a database that contains all relevant information about the hardware and software components used in an organization’s IT services and the relationships between those components.” The information stored in the CMDB is important because it allows the entire organization to refer to technical components in the same manner. In a larger organization, the team that is responsible for provisioning hardware and software components will often be responsible for also making sure that any information related to newly provisioned components makes its way into the CMDB. There is often an administrator or set of administrators that will maintain the information in the CMDB. The data will then be consumed by a large number of teams, including IT Support, Project Teams, and Finance.

When the accuracy of the data is not complete, the teams consuming the data do not have the ability to speak the same language regarding IT components. The Finance Team may allocate dollars based on the number of components or breakdown of types of components. If they do not have adequate information, they may fail to allocate the right budget for the project teams to complete their work on time. A different understanding of enterprise components may cause delays in assistance from the IT Support organization, which has the potential to push out timelines and delay projects.

One potential solution to this issue: make one team responsible for maintaining the accuracy of the data from collection to consumption. As mentioned before, this team needs to have an understanding of where the data comes from, how it is stored, how it is consumed, and the ability to recognize when the data is not accurate and the steps required to correct the information. The data integrity team must be accessible to the rest of the organization to correct data accuracy problems when they arise. As the team grows and matures, they should target developing proactive measures to test that data is accurate and complete so that they can solve data integrity issues before they impact the user. By assigning specific ownership over the entire data lifecycle to one team, the organization can enforce accountability and integrity and mitigate the risk that leaders make poor decisions based on false information.

Links:

[1] Digital Guardian: https://digitalguardian.com/blog/what-data-integrity-data-protection-101
[2] Investopedia: https://www.investopedia.com/articles/professionals/120115/data-integrity-analyst-job-description-average-salary.asp
[3] SearchDataCenter: https://searchdatacenter.techtarget.com/definition/configuration-management-database

Using Social Media to Screen Job Candidates: Ethical and Future Implications
By Anonymous | June 24, 2019

Image Source: https://www.cfm-online.com/research-blog/2017/7/26/eu-looks-at-limiting-employer-social-media-snooping

Hiring qualified people is hard. Most of the time, the foundation of a hiring manager’s decision is built off of a 1-page resume, a biased reference or two (sometimes none), and a few hours of interviews with the candidate on their best behavior.

It’s no surprise that around 70% of employers have admitted to snooping on personal social media sites as a method for screening candidates [1]. Since hiring someone who isn’t the right fit can be expensive, it’s only natural for companies to turn to Facebook, Twitter, Instagram, or other social media sites to get a deeper glimpse into the personality they’re hiring. Unfortunately, there’s a lot that can go wrong for all parties involved due to the ethical implications.

What could go wrong?

Using social media to screen candidates doesn’t just weed out people who are vocal online about their criminal or illegal behavior. Doing this can lead to hiring managers screening out perfectly qualified candidates.

Recently, CIPD (an employee advocate group based in London) wrote a comprehensive pre-employment guide for organizations to follow, and included a section on using social media for job screening [2]. They outlined the risks of employers doing this, which included a case study about a company deciding not to hire a transgender candidate, even after indicating that the individual was suitable for the job prior to the social media check. This was considered an act direct discrimination under a protected characteristic, brought on by the company using social media to get more information on the candidate.

It doesn’t stop there. For some people, it’s common sense that employers review social media profiles, and they are able to keep their private thoughts secured. However, not everybody is a social media expert, and deciphering exactly is and isn’t private can be unwieldy. Many people are not aware that they are consenting to disclosing posts from 5+ years ago to potential employers. When companies don’t directly disclose that all content from personal social media sites are subject to review, this could be considered a breach of privacy for individuals who are unaware.

The Future of Social Media Screening

Manually reading through social media sites for potential issues with the candidate is time consuming. Why can’t someone your just create an algorithm that parses through social media content when it’s available, and labels attributes of your employees for you?

Image Source: http://fortune.com/2017/05/19/ai-changing-jobs-hiring-recruiting/

With the massive influx of artificial intelligence being leveraged within the job-hunting industry, it’s surprising that this isn’t already an industry norm. However, there are a myriad of potential ethical concerns around creating algorithms to do this.

It’s entirely possible that job candidates can fall victim to algorithmic bias, and be categorized as something they’re not because of an unperfected algorithm. If someone is new to social media and undergoes a screening like this, it’s possible the result will find no positive traits for the candidate, and the company will reject the candidate based on the algorithm’s decision.

Between the start-ups that continue to sprout up for the purpose of data mining to gain valuable insights on individuals and the “Social Credit Score” going live in Chine in 2020 [3], it’s hard to discount the possibility of algorithmic social media screenings that score how “hirable” a candidate is becoming prevalent. Because of this, all aspects of the hiring process should continually be subjected to ethical laws and frameworks to protect job candidates from unfair discrimination.

References

[1] https://www.cbia.com/news/hr-safety/employers-continue-rejecting-jobseekers-social-media/

[2] https://www.cipd.co.uk/knowledge/fundamentals/emp-law/recruitment/pre-employment-checks-guide

[3] https://digitalcommons.law.yale.edu/cgi/viewcontent.cgi?article=1122&context=yjolt

Ethical Implication of Generative AI
By Gabriel Hudson | April 1, 2019

Generative data models are rapidly growing in popularity and sophistication in the world of artificial intelligence (AI). Rather than using existing data to classify an individual or predict some aspect of a dataset these models actually generate new content. Recently developments in generative data modeling have begun to blur lines not only between real and fake, but also between machine and human generated content creating a need to look at the ethical issues that arise as the technologies evolve.

Bots
Bots are an older technology that has already been used over a large range of functions such as automated customer service or directed personal advertising. Bots are generative (almost exclusively creating language), but historically have been very narrow in function and limited to small interaction on a specified topic. In May of 2018 Google debuted a Bot system called Duplex that was able to successfully “fool” a significant number of test subjects while carrying out daily tasks such as booking restaurant reservations and making a hair salon appointment (link). This, combined with ubiquity of digital assistants, sparked a resurgence in bot advancement.

Deepfake
In this case Deepfake is a generalized term used to describe very realistic “media” (such images, videos, music, and speech) created with an AI technology know as a Generative Adversarial Network (GAN). GANs were originally introduced in 2014 but came into prominence when a new training method was published in 2018. GANs represent the technology behind seemingly innocuous generated media such as the first piece of AI generated art sold (link):

as well as a much more harmful set of false pornographic videos created using celebrities faces
(link):

The key technologies in this area were fully release fully released to the public upon their completion.

Open AI’s GPT-2
In February 2019 Open AI (a non-profit AI research organization founded in part by Elon Musk) released a report claiming a significant technology breakthrough in generating human sounding text as well as promising sample results (link). Open AI, however, against longstanding trends in the field and their own history chose not to release the full model citing potential for misuse on a large scale. Similar to GPT-2, there have also been breakthroughs in generative technology in other media like images, that have been released to the public. All of the images in the subsequent frame were generated with technology developed by Nvidia.

In limiting access to a new technology Open AI brought to the forefront some discussions about how the rapid evolution of generative models must be handled. Now that almost indistinguishable “false” content can be generated in large volume with ease it is important to consider who is tasked with deciding and maintaining the integrity of online content. In the near future, discussions must be extended about the reality of the responsibilities of both consumers and distributors of data and the way their “rights” to know fact from fiction and human from machine may be changing.