Baby Bonds: An Investment In Our Future

Since the 2008 financial crisis, income and wealth inequality has become a prominent subtext of our national conversation. Politicians, ranging from President Obama and local city governments, as well as social movements, including Occupy Wall Street and the Fight For $15 (which has successfully fought for a $15 minimum wage in several dozen cities and states), have drawn attention to the nation’s highly uneven private monetary distribution.

Considering that in 2014, the top 0.1% of earners brought in 184 times the income of the lower 90% of Americans combined, while members of the Forbes 400 (the 400 wealthiest Americans, or just over 0.000001% of the population) held greater aggregate wealth than the bottom 61% of Americans, this debate is unlikely to cease anytime soon.

A range of potential solutions have been floated, including raising capital gains taxes (assessed on various types of investments), increasing the income tax on high earners, and reforming the estate tax (levied on inheritances of a certain size), to dampen the effects of the intergenerational transfer of wealth.

Each of these approaches is focused on altering the post-earnings distribution of income and wealth, through the use of taxes. There’s nothing per se wrong with that method; after all, tax policy is one of the primary mechanisms through which government can influence economics.

Yet, what if policymakers try to tackle this issue from a different angle, by working to shift the pre-tax distribution of income and wealth? Such a strategy would involve crafting solutions which empower those from families who make less money, and sit at the lower end of the wealth spectrum, to increase their incomes, and build wealth.

In the past several years, studies have shown that economic mobility, in terms of income, has not changed much over the past several decades. That is, the chances of moving up to the top 20% of earners as an adult, given a childhood background where one’s family was in the bottom 20%, has remained relatively consistent. Yet, the numerical odds are still rather low: only 8.4 to 9% of Americans will make this jump.

Much of this has to do with educational achievement. There is a strong correlation between education and earnings, yet just 5% of Americans whose parents did not finish high school (and are thus more likely to be clustered amongst the poorest 20% of the population), complete college. More broadly, of those aged 25–34 (in 2014), only 20% of men, and 27% of women, achieved a greater level of education than their parents.

An obvious reason for this unfortunate reality, is that children born into poverty face a range of disadvantages before they even begin school, including lower exposure to vocabulary and conversation, to malnutrition, and the presence of violence and drug use. These children often attend schools where fellow pupils also hail from socioeconomically challenged backgrounds, teacher quality tends to be lower, and an inequitable distribution of public funds leads to a lack of other resources.

As if all of this weren’t enough, those who do make it to college often face rising educational costs. A landmark 2007 study found that the impact of family income on college attendance, grew markedly from the early 1980’s to the early 2000’s. Why? In part, the study suggested, rising educational costs are harming those who are “borrowing constrained”, that is, less affluent students, who often must accept the maximum amount allowed from student loan programs (sometimes while facing other financial pressures, like assisting their families), also lack other sources of funding.This is an unfortunate Catch-22, since higher education is often required in order to escape poverty, yet, it often simply costs too much.

Now, let’s consider the wealth gap. The wealthiest 20% of American households hold just under 90% of all wealth in the nation, while the bottom 40% of households actually have negative wealth, that is, they actually owe more than the total value of their assets, i.e., such households are in debt.

For those on the lower end,of the wealth curve, breaking out of this cycle is even more challenging than outearning their parents (i.e.narrowing income inequality). Since so much wealth in the United States is transferred intergenerationally, even if someone from a poor background completes advanced schooling, and earns far more than his or her parents, he or she is already at a substantial disadvantage, compared to peers from wealthier families, whose families might have helped pay for school, provided the down payment for a home, or helped fund a new business. This chasm remains for generations to come.

Yet, wherever we face major challenges, solutions are never too far away. One of the most fascinating approaches to bridging the wealth gap, is the idea of “baby bonds.” This phrase, coined by late historian Manning Marable, came into prominence thanks to a 2010 paper, published by economists Darrick Hamilton of The New School and William Darity Jr. of Duke University.

Hamilton and Darity’s work sought to address racial disparities in the distribution of wealth. According to a 2013 study from the Institute for Policy Studies, the average white household was almost 7.7 times as wealthy as it’s African-American counterpart, and had almost 6.7 times the wealth of the average Latino household.

Why is this? Hamilton and Darity cite to a study by economists Maury Gittleman and Edward Wolff, which found that, once controlled for income, African-Americans and whites have similar savings rates, eliminating one source of this wealth disparity. However, inheritances play a large role in raising the wealth levels of whites, relative to African-Americans. This effect is so pronounced that the median wealth of an African-American household headed by a college graduate, is lower than that of a white family whose primary earner dropped out of high school. Hamilton and Darity also note that lending and housing policies have been disproportionately harmful to African-Americans, which has an added impact on wealth creation.

Thus far, we have established that the wealth gap in the United States is direly large, stems largely from the passage of wealth through inheritance, and has a significant racial component. So, what does all of this have to do with baby bonds?

As Hamilton and Darity see it, the “post-racial” narrative, which argues that racism and discrimination are no longer major factors in holding back African-Americans, has become increasingly prevalent, particularly since the election of President Obama. Under this view, African-Americans must assume greater personal responsibility for their social and economic conditions, rather than looking to the larger society, or the federal government. Advocates of post-racial politics also believe that African-Americans ought to focus on supporting programs designed to uplift Americans of all races, rather than racially focused remedies.

While Hamilton and Darity aren’t swayed by claims of post-racialism, they seem to acknowledge that these arguments have gained considerable traction, such that there is unlikely to be much support for “race specific social policies” to bridge the racial wealth gap, or other disparities.

Here’s where baby bonds offer a solution. Around 1990, a few small pilot programs were established to narrow the wealth gap. Amongst these was the Savings for Education, Entrepreneurship and Down-payment (SEED) initiative, which funds Children’s Development Accounts (CDA), basically, savings accounts to help children build wealth, starting at birth.

Hamilton and Darity argue for widespread implementation of a program like SEED. Under their approach, around 75% of American newborns would be allocated an initial principal amount, based on familial wealth (rather than income). Children born into the least wealthy households would recieve a larger allocation (perhaps $50,000 to $60,000), while those from wealthier families, who are capable of passing on greater wealth, would recieve progressively smaller amounts (babies from the wealthiest households would recieve no baby bonds, given their not-insubstantial existing wealth). Overall, Hamilton and Darity foresee around 3/4 of children born every year, participating in this program.

These funds would be placed in federally managed accounts, with a guaranteed annual return of 1.5% to 2%, such that by the time a child from one of the poorest families turns 18, he or she would have $78,000. This money would then be directed towards a “clearly defined asset enhancing activity,” that is, an avenue which will empower recipients to build wealth. Examples offered by Hamilton include financing (partially) debt-free higher education, investing in a business, or purchasing a home (given housing costs in much of the country, probably by just providing a down payment).

While bond holders have their choice of how exactly to allocate these funds, monies must be spent on some sort of activity which builds long-term wealth. As Hamilton noted, in many poorer families, relatives often need financial assistance, to cover various expenses. Such uses, however commendable, will not be permitted, since these funds won’t strengthen the long-term finances of bond beneficiaries. Additionally, vigorous regulations would be implemented, to ensure that beneficiaries are not defrauded by those who see a young person looking to improve his or her life, as an easy target for theft.

Of course, whenever a proposal for government spending is offered, the inevitable question arises: How will we pay for this? In their paper, Hamilton and Darity projected that the baby bonds program would cost about $60 billion per year. Assuming costs have stayed fairly consistent (a reasonable assumption, given birth rates), this amount is just over 10% of the yearly allocation to the Department of Defense (which is already rife with bureaucratic waste) . It is also less than the considerable tax revenue lost yearly to the mortgage interest deduction, a largely ineffective. policy. Funding baby bonds is possible, if lawmakers are willing to make some reasonable fiscal adjustments, for a program that holds considerable promise.

Ultimately, it will be possible to finally lessen a yawning wealth gap, which stems in major part from decades of wealth transfer between generations, rather than uniquely meritorious behavior, of those who possess (often, inherit) wealth . Baby bonds can directly impact the college completion rates of poorer Americans (recall the aforementioned discussion of borrowing constraints), given that promising students (based on test scores) of lesser means, complete college at about the same rate as wealthy youngsters with considerably weaker academic skills.

What’s more, those baby bond recipients who do complete college, will be graduating with a lighter debt load, and greater earnings potential, than their counterparts who did not finish school. This will not only reduce income inequality, but allow baby bond program participants to build (and bequeath) wealth at a faster pace. Over time, this virtuous cycle will (partially) offset unearned differences in wealth.

Baby bonds are not a panacea for the complex challenges we face, in terms of wealth and income inequality. Changes to the tax code, better K-12 schools, strong antipoverty efforts, improved financial education, can all play important roles in overcoming these challenges. What’s more, in an era where technological innovation (specifically, artificial intelligence, machine learning, deep learning and robotics), will replace lots of blue collar as well as higher skill jobs, inequality will likely remain an issue for decades to come.

Yet, we must make an earnest effort to build a fairer economic system, one where the birth lottery does not play such a large role in determining one’s income and wealth, not to mention that of future generations. Baby bonds are a clear step in that direction.

Smart Robots And The Service Revolution

                                                   Photo Credit: Australian Financial Review

In June 2015, the Japanese company SoftBank Robotics Corp. released a product named Pepper, which sold out in a mere 1 minute. What’s so special about Pepper? Pepper is a robot, standing about four feet tall, who can read and process a range of human emotions, in order to understand and communicate with people. Pepper can “feel” relaxed, anxious, happy, scared and much more. Because Pepper is connected to a cloud-based system, and utilizes machine learning, each Pepper robot will gradually become more adept at social engagement, and learn from the collective range of interactions, of all of his counterparts.

At first glance, Pepper might appear to be little more than a charming novelty. Yet, SoftBank’s invention is an early act in a long-heralded technological revolution, which will transform our world forever. In the coming years, an ever-growing array of socially and emotionally intelligent robots will take on a wide range of customer-facing functions, allowing businesses, governments, and nonprofit organizations, to deliver more effective services, to a greater range of people, at a much lower cost.

Developing software which replicates human behavioral patterns and interaction, is hardly a new goal. After all, Alan Turing’s eponymous Turing Test dates back to 1950. In the decades since the publication of his visionary paper “Computing Machinery and Intelligence” we’ve inched closer and closer to meeting his bar for artificial intelligence (AI), that is, a machine which is humanesque enough that, during a normal conversation (written, rather than verbal), the average person can’t tell whether he or she is interacting with a live human being, or a robot.

In just the past 7 years, we’ve witnessed the rise of Siri on the iOS platform, watched Google’s AlphaGo beat the world’ greatest Go player, and been awed as IBM’s Watson defeated some of the most skilled Jeopardy contestants of all time. Deep-learning based image recognition and language translation programs have improved at a remarkable pace, while self-driving cars inch closer and closer to actual market implementation.

The trend is clear. Software-based machines, utilizing the power of artificial intelligence, deep learning, and machine learning, can learn to perform a growing array of tasks, often, better than humans can. This trend will become only more pronounced over time.

Most of the examples offered above, however, don’t require much in the way of social or emotional intelligence. After all, a robot playing Go, or competing in Jeopardy, needn’t form a meaningful personal connection, with either an audience, or users. Yet, for smart robots to realize their full potential, they will need to do just that; that is, understand human communication styles, conversations, and emotional states, and engage accordingly.

At this very moment, teams of determined technologists are working to achieve this goal. Companies such as Affectiva, a spinoff of the M.I.T. Media Lab, have developed software to identify and understand (at a deep level) people’s emotional states, largely by analyzing facial expressions. Affectiva’s tools have applications ranging from empowering individuals with autism, to more intelligent advertising, and smarter videoconferencing. As Affectiva collects and learns from more and more human emotional data, the company’s products will become increasingly effective.

Of course, much of today’s communication isn’t done in person, or by video, but rather, on the phone. This is particularly true in sales and customer service. Cogito, another child of the M.I.T. Media Lab, has developed software to guide phone-based customer support employees in providing more effective customer support. Cogito does this by tracking the speed of conversations, pauses, tone, interruptions and more, and providing feedback, based on comparison to typical conversations within various situations and industries.

Cogito is currently working with cadets at West Point to improve negotiation skills, and with employees at Aetna and BlueCross to reduce customer call time, as well as callbacks. What’s particularly fascinating about Cogito, is that it combines both qualitative and quantitative data, to allow for rapid improvement and feedback.

Firms like Cogito and Affectiva are taking hold at a time when chatbots, basically, messaging tools which communicate with humans through the use of artificial intelligence and natural language processing, are growing in popularity. Chatbots are still in a relatively early stage of implementation, but are being used by major companies like Bank of America ,H&M and Whole Foods, to engage with customers more effectively. At the same time, a variety of independent chatbot platforms have started to take root, ranging from personal assistants to customer support tools. Each of these devices has the potential to both complement, and in some cases supplant, human involvement.

Taken in aggregate, we can sketch out a roadmap of where these technological advancements might lead us. Innovations offered by firms like Affectiva and Cogito, have allowed software to gain a deeper understanding of human emotional states, and communication styles. Right now, many of these insights are utilized to guide humans towards more effective communication; such as the use of Affectiva’s software by advertising executives, or Cogito’s implementation at West Point.

Eventually, however, the knowledge contained in these tools might be integrated directly into various devices. Imagine a digital chatbot, or perhaps a video or voice-enabled robot, which is equipped with the sort of understanding of conversational dynamics offered by Cogito, paired with the emotional intelligence by Affectiva, and other insights from neuroscience, psychology, and other disciplines. What’s more, what if this chatbot can become progressively more intelligent, by engaging in the sort of deep learning that’s increasingly driving so much of today’s most impactful innovations? How might we utilize such an innovation?

Simply put, we would see a transformation in service-focused operations, whether at multibillion dollar public corporations, startups, nonprofits, and everything in between.

Let’s consider large firms with a substantial customer service component, such as banks, credit card companies, and wireless service providers. If they were able to shift over to a robot-based approach, the need for a large (human) support staff will be eliminated. A robot could perform much of the work of humans, at a mere fraction of the cost. What’s more, thanks to techniques like machine learning, over a certain period of time, these robots will become increasingly effective at serving customers. Soon enough, customer service will become less of a cost center, as companies use humans for only the most complex cases. As noted earlier, these developments are already taking place at financial institutions, and will surely be adopted elsewhere, in the near future.

For startups and nonprofits, focused on solving big problems, the potential impact is even greater. Picture a small educational technology company, seeking to solve a complex problem, say, helping students become more skilled at math. What if such a company were to create a robot, similar in functionality to those we considered earlier, which provides students with individualized instruction and feedback? A student might work through a math problem at home, displaying his or her work for review by a robot, which could then offer a range of valuable guidance, tailored to her or his specific learning style, and curricular needs. Just as with customer service robots, these robots would grow smarter, and more effective, over time.

Utilizing this approach, an intrepid educational technology company, made up of a relatively small staff, could eventually reach millions of students, and at least partially alleviate gaps in educational outcomes, due to parental income and education levels, as well as teacher quality, all factors which hold back so many students. This certainly isn’t a replacement for effective, engaged public education, or parental involvement, but it can serve as a highly useful, necessary supplement.

For nonprofits and government, robots can help make some rather scarce services, considerably more accessible. Legal aid is a good example. Today, the vast majority of defendants in certain types of civil court matters, such as debt collection and housing rights, are not represented by an attorney. This is largely due to a combination of defendants having lower incomes, while legal services organizations face a lack of funding.

With a chat robot, an unrepresented litigant could ask questions, and analyze and understand documents, as well as the broader legal process, surrounding his or her case. This robot would respond with individualized guidance, helping to simplify and level the legal playing field, at a relatively low cost. Such tools are already being utilized to assist criminal defendants in the UK. It is worth noting, however, rules around unauthorized practice of law would have to be reconciled with robot-based legal assistance.

It’s easy to imagine similar applications, across a variety of other governmental and nonprofit arenas. Everything from tax preparation and construction permits, to social services and healthcare, could be made more transparent and accessible, with the rise of chatbots and similar interfaces.

We can’t overlook the reality that robots have not yet surpassed raw human intelligence (though some experts do expect them to do so within the next 100 years). Neither have robots achieved human levels of emotional intelligence (though experts in the field, most notably Google’s resident futurist, Ray Kurzweil, believe this will happen before 2030). The best robot-focused service models, particularly today, as we await further advancements, will also require humans, to step in where robots fall short. Therefore, the robot revolution will be a fast-moving, yet evolutionary process.

Some might raise another objection: user adoption. That is, will those being asked to interact with robots (say, a customer of a bank, or a client of a legal aid organization), be comfortable with and open to working with some sort of robot? After all, this is a novel means of interaction. Yet, the history of technology shows us that humans will eventually alter a range of behaviors, when presented with a better option. People once used to travel by buggy and ship, but eventually chose the convenience of cars and airplanes. More recently, we shifted from using typewriters to personal computers, while searching for information and communicating online, sharing a range of information on Facebook and Twitter, and messaging each other on our mobile phones. For comfort, convenience, and savings, people are willing to alter their conduct.

Ultimately, for those young businesses with a substantial service component, smart robots will allow them to grow and scale more quickly, which would be more challenging if forced to hire lots of human representatives. For larger companies, whose profits are often under pressure, and in a time when corporate lifespans are decreasing, a notable cost center is reduced. Of course, for firms of all sizes, in the era of the social media and online reviews, where reputations are increasingly transparent, effective efficient robot-based service will only bolster goodwill for the purchasing public.

Nonprofits and governmental organizations, who must often grapple with uncertain funding, along with an ever-present demand for their services, robots will help them to serve more of those who are most in need, also at a lower cost. The societal effects of this increased access to vital services, can only be positive.

A range of smart robots are going to change our lives immeasurably. Get in, buckle up, and really, enjoy the ride.

Algorithms, Big Data And Accountability

                                       Credit: www.ug.ru

Everywhere you look, you’ll find them. In trouble with the law, and facing possible jail time? She is an ever-present fixture in the judge’s chambers, and you can bet we’ll hear plenty from her. Want to obtain a car loan? You can see him sitting on the loan officer’s desk, waiting to offer his thoughts. What about when you apply for a job? Quite often, she is the ultimate gatekeeper, parsing through every single resume, and deciding who will or won’t get a call back.

We live in an era where large sets of data (often referred to as Big Data), and the formulas used to analyze this information coherently (algorithms), are used to make highly impactful decisions, affecting almost every facet of our lives. The role of these tools in our lives will continue to grow.

However, sometimes these algorithms are of highly questionable accuracy, and use flawed or incomplete data, to reach decisions which affect millions of lives, at times causing considerable harm. What’s more, despite their outsized roles in our lives, many of these algorithms remain secret, with their inner workings unknown to everyone except for their creators.

Clearly, this situation cries out for greater transparency, and more effective regulation. This can be accomplished in the form of a federal agency, similar to the FCC, which would police the implementation of those specific algorithms which can have a significant negative impact on a large swath of the American public.

ProPublica recently published a detailed piece on how algorithms known as risk assessments, are used to determine a criminal defendant’s risk of reoffending. The results of these calculations are then factored into decisions concerning bail, sentencing, parole and more. ProPublica's study of more than 7,000 individuals arrested in Broward County, Florida, uncovered some rather troubling flaws in the risk assessment model developed by Northpointe,a private software company whose secret algorithms are utilized in jurisdictions throughout the nation.

Northpointe’s algorithms are based on questionnaires answered by defendants, which probes a range of data, including life prior to arrest, by asking questions about everything ranging from the arrest records and criminal history of a defendant’s family and friends, academic track record, personality traits, and drug and alcohol usage.

These models have proven quite unreliable in predicting whether someone who was arrested, will actually commit a violent crime in the future. In a two year period following an arrest (the same benchmark used by Northpointe’s creators in designing their software), just 20% of those who were thought likely to commit a violent crime, actually did so (the recidivism rate for all crimes, at a 61% accuracy rate, was “somewhat more accurate than a coin flip”).

Under Northpointe’s model, African-Americans were almost twice as likely as whites to be wrongly labeled as likely to re-offend (that is, Northpointe’s model incorrectly predicted future criminal behavior for African-American defendants, at twice the rate of whites). Additionally, whites were considerably more likely than African-Americans to be labeled as at low risk of recidivism, and yet end up in prison again (i.e. Northpointe’s algorithm majorly underestimated recidivism risk amongst whites) .

ProPublica’s researchers wondered if this racial gap might be due to other factors, including prior criminal history, age and gender. However, the disparities stubbornly persist, even when controlling for those variables. While Northpointe disagrees with ProPublica’s findings; since the actual algorithm remains secret, there’s little way to publicly debate and assess it’s functionality.

In some jurisdictions, Northpointe’s algorithms allow judges to decide whether prisoners should be granted pretrial release, or directed towards a rehabilitation program. In other places, like La Crosse County, Wisconsin, judges have utilized risk assessment scores to identify certain individuals as being at high risk for reoffending, and handed down longer sentences as a result (even Tim Brennan, the statistician who cofounded Northpointe, has expressed opposition to using this software for sentencing ).

The use of secretive, “black box” algorithms in the criminal justice system also poses major constitutional concerns. Under the Due Process clause of the Constitution, before depriving a person of his or her rights, the government must apply processes that are, at their core, fair and non-arbitrary. How does one challenge the (often inaccurate) statistical recommendations of a computerized formula, when we can’t even know how it’s findings were in fact derived? Secretive algorithms pose a unique problem to protection of this core right.

Credit scores and data, specifically those generated by FICO, and the three major credit bureaus (Experian, Equifax, and TransUnion), raise substantial concerns of fairness and transparency. FICO scores, which are the most widely used measure of creditworthiness by lenders, employers, and others, are made up of a mix of one’s payment history, debt load, the type of credit used, and other data found in credit reports.

While we have a basic idea of which ingredients go into a FICO score, as the Fair Isaac Corporation (the parent company of FICO) acknowledges: “The importance of any one factor in your credit score calculation depends on the overall information in your credit report.....therefore, it’s impossible to measure the exact impact of a single factor in how your credit score is calculated, without looking at your entire report.” Thus, FICO (as well as the VantageScore product, created by the three credit bureaus), offers little real clarity, in terms of allowing an individual to figure out why a credit score falls exactly where it does. Changes in various data points on a credit report, can affect FICO scores in a somewhat unpredictable manner.

When one considers the very substantial impact of a credit score, this opacity and secrecy is especially troubling. FICO numbers typically play a major role in determining the interest rate one pays on a mortgage or car loan, as well as whether one is able to rent an apartment, and, in some states, be hired for a new job.

While the Fair Credit Reporting Act provides methods for customers to challenge and remove inaccurate information from credit reports (which is crucial, considering an FTC study found 25% of all credit reports contain at least one material error), there’s very little that any individual customer can do to challenge the secretive verdict handed down by FICO. If there is some sort of fundamental flaw in how FICO assesses credit, such that FICO scores are a partially inaccurate gauge of creditworthiness, there is no way to know, because just like Northpointe’s software, this pivotal algorithm remains secret, hidden from outside scrutiny.

Securing gainful employment is one of the most critical (and sometimes challenging) aspects of our lives. Here too, secret algorithms play an increasingly prominent (and rather troubling) role. With around 72% of all resumes never initially reviewed by human eyes, but rather through computer programs, applicants who are skilled in sprinkling buzz phrases and keywords throughout their resume, are often favored in hiring. Job-matching algorithms which assess the likelihood of employee retention and success, can be further biased against those who are poor, as Xerox discovered with a now-defunct program they used for evaluating applicants, based on the likelihood of an employee quitting his or her job .

Applicants are also often asked to take computerized personality and cognitive tests (again, based on private algorithms and data sets), which offer questionable predictive value of employee performance, but can be used to illegally exclude those with disabilities, or individuals whose evaluations fall outside of some desired bandwidth (litigation on the legality of these practices is ongoing). With such tests being used to evaluate 60 to 70% of job applicants in the United States, these assessments have a large impact on hiring practices.

What conclusions can we draw from all of this? Are algorithms inherently unfair and prejudiced? Not quite. However, the process surrounding the implementation of these high-impact tools, clearly requires some significant changes.

Morris Hardt, a research scientist at Google, has detailed several sources of algorithmic unfairness. First, he notes, machine learning (that is, algorithms which behave intelligently, and learn from the data they are provided with), will typically reflect the patterns found in data; that is, if there is a “social bias” against any group of people, the algorithm is likely to pick up on and mimic such a pattern (Hardt cites the work of Solon Barocas and Andrew Selbst, who found that algorithms can  “inherit the prejudices of prior decision makers...in other cases, data may simply reflect the biases that persist in society at large.”). Thus, the inherent unfairness of the criminal justice system, or the employee hiring process, towards certain individuals or groups, will be reflected in algorithms which address these arenas.   

Hardt also points out that in assessing data, samples of data concerning underrepresented or disadvantaged groups, are by necessity smaller, and thus less representative, than for the general population. After all, if the premise of big data is that more data can improve predictive value, then less data often result in weaker predictions.

Beyond these issues with data, there is another major issue surrounding the use of algorithms: transparency. Since the mechanics of so many algorithms with a large public impact remain completely secret, we often don’t know whether they are working fairly, or properly.

Are we truly confident that Propublica’s findings regarding the flaws in Northpointe’s recidivism predictions are some sort of anomaly, rather than the norm, in the world of criminal justice algorithms? How certain are we that those formulas which assess the personalities of job applicants, are a fair and accurate reflection of whether a company should consider hiring someone? Do FICO scores, and the (often erroneous) data on which they are based, provide a reasonable snapshot of a borrower’s likelihood of repaying a loan? And if so, could they be made even better?

Fortunately, there are several concrete steps we can take, to overcome this problem. First, we need to carefully define which sorts of algorithms we should be most concerned about. After all, Big Data, and it’s associated algorithms, are used for undertakings ranging from cancer treatment, trading by hedge funds, threat assessments by the US military, and countless other applications in so many fields.

So how do we decide which algorithms ought to be subject to greater scrutiny? Cathy O’Neill, a mathematician and data scientist who recently published an acclaimed book warning of the dangers of algorithms and Big Data, offers a three part test to answer this question. First, is an algorithm high impact, that is, does it affect a large number of people, and carry major consequences for their lives? (Those pertaining to jobs and criminal justice are two examples O’Neil cites). Second, is it algorithm opaque; that is, people who are assessed by these formulas, don’t actually know how their scores are computed (all the examples we have considered thus far meet this criteria). Third, is an algorithm in fact destructive, that is, can it have a major negative impact on a person’s life (again, the aforementioned issues all seem to fit this test)?

What specific steps can we take to limit the potential for algorithms and Big Data to inflict harm? We need to develop rigorous due process and appeals procedures. One promising solution, which was recently implemented by the European Union (taking effect in 2018), requires that any decision based “solely on automated processing” which includes “legal effects” or “similarly significantly affects” an individual, be subject to “suitable safeguards,” including an opportunity to obtain an explanation of an algorithmic decision, and to challenge such decisions.

Here in the United States, comparable legislation, ideally passed at the federal level, is greatly needed. Such a law would first apply O’Neill’s test, to determine whether an algorithmic process warrants greater scrutiny. If it does, then a regulatory body, much like the Federal Communications Commission (FCC), ought to be tasked with providing oversight. Let’s call it the Algorithmic And Data Implementation Commission (AADIC).

How might the AADIC fulfill this mission? Just as with the FCC, a group of commissioners, appointed by the president, and confirmed by Congress, would play a primary role in offering policy guidelines for algorithmic processes generally (analogous to what the FCC did in formulating “net neutrality” rules), and helping determine whether a particular algorithm produces decisions that are fair, accurate and representative. Ideally, at least some of these commissioners would have backgrounds (both academic and commercial)  in fields like data science, statistics, and more generally, the collection and processing of large data sets.

In deliberating on and reaching such decisions, AADIC commissioners (and the public) will be provided with both the underlying formulas, as well as a sample of the data utilized by these algorithms. Commissioners would solicit public comment on the algorithms and data, from both those in support of and opposed to, a particular sort of decisionmaking (similar to amicus briefs to the Supreme Court). Of course, it is crucial for the AADIC to also recieve trusted, impartial advice. Towards this end, the commission would retain it’s own staff of experts, who could assess the effectiveness and overall performance of any data and algorithm sets.

If a majority of commissioners decides that a particular use of algorithms was somehow flawed or problematic, they can veto it’s use for public purposes, and send it back to it’s creators for further improvement and revision. Of course, just as with any government agency, the AADIC requires checks on its’ powers. Just like the FCC, decisions of the AADIC will be challengeable in federal court.

In an era where trust in the federal government is weaker than ever, many will be understandably skeptical of expanding the federal government’s regulatory authority, into yet another sphere.  I too am wary of the gargantuan bureaucracy we find in Washington DC, and certainly don’t see government as a panacea for all the challenges we face. Also, neither the AADIC, nor any other governmental body, to become a crippling roadblock for progress and innovation.

With that said, in this instance, the state must play a prominent role. The scope of algorithmic and data-based decisions in our lives continues to grow unabated, and is in dire need of some rigorous safeguards. While states, public interest groups, and private citizens can all play a positive role here, the authority of the federal government is key to offering the neccessary degree of coordination, oversight, and enforcement, to facilitate fairness, and reduce abuses. In this sense, some algorithms are no different from prescription drugs or securities.

The myriad new possibilities opened up by advancements in Big Data, and algorithmic processes, is nothing short of incredible. From insurance to healthcare to law to transportation, and so many other fields, these tools are remaking entire industries, and bringing an unprecedented degree of insight, efficiency, and cost reduction to our lives. Yet, as we now know, these tools can also be used in a harmful manner, and we must guard against such abuses. The AADIC is a decisive step in that direction.

  

  ,