AI Policy Nuggets II

What Can Biden Do for Science?

A Science|Business Webcast presented a forum of public and private sector leaders discussing ideas about the need for the president-elect to convene world leaders to re-establish ‘rules of engagement’ on science.

Brookings Webinar on the Future of AI

“On November 17, 2020, the Brookings Institution Center for Technology Innovation hosted a webinar to discuss the future of AI, how it is being deployed, and the policy and legal issues being raised. Speakers explored ways to mitigate possible concerns and how to move forward safely, securely, and in a manner consistent with human values.”

Section 230 Update

Politico reports that “Trump for months has urged Congress to revoke industry legal shield Section 230, while its staunchest critics largely pushed to revamp it instead. But the president’s more drastic call for a total repeal — echoed by Biden for very different reasons — is gaining traction among Republicans in Washington. The NYT reported Thursday that White House chief of staff Mark Meadows has even offered Trump’s support for a must-pass annual defense spending bill if it includes such a repeal.”

The European AI Policy Conference

AI may be the most important digital innovation technology transforming industries around the world.
“Businesses in Europe are at the forefront of some of the latest advancements in the field, and European universities are home to the greatest concentration of AI researchers in the world. Every week, new case studies emerge showing the potential opportunities that can arise from greater use of the technology.” The European AI Policy Conference brings together leading voices in AI from to discuss why European success in AI is important, how the EU compares to other world leaders today, and what steps European policymakers should take to be more competitive in AI. “The European AI Policy Conference is a high-level forum to connect stakeholders working to promote AI in Europe, showcase advances in AI, and promote AI policies supporting its development to EU policymakers and thought leaders.”

Policy Issues from AI and Ethics

The inaugural issue of the new journal AI and Ethics contains several articles relevant to AI and Public Policy.

Jelinek, T., Wallach, W. & Kerimi, D. “Policy brief: the creation of a G20 coordinating committee for the governance of artificial intelligence” AI Ethics (2020). https://doi.org/10.1007/s43681-020-00019-y

This policy brief proposes a group of twenty (G20) coordinating committee for the governance of artificial intelligence (CCGAI) to plan and coordinate on a multilateral level the mitigation of AI risks. The G20 is the appropriate regime complex for such a metagovernance mechanism, given the involvement of the largest economies and their highest political representatives.

Gambelin, O. “Brave: what it means to be an AI Ethicist” AI Ethics (2020). https://doi.org/10.1007/s43681-020-00020-5

This piece offers a preliminary definition of what it means to be an AI Ethicist, first examining the concept of an ethicist in the context of artificial intelligence, followed by exploring what responsibilities are added to the role in industry specifically, and ending on the fundamental characteristic that underlies it all: bravery.

Smith, P., Smith, L. “Artificial intelligence and disability: too much promise, yet too little substance?” AI Ethics (2020). https://doi.org/10.1007/s43681-020-00004-5

Much has been written about the potential of artificial intelligence (AI) to support, and even transform, the lives of disabled people. Many individuals are benefiting, but what are the true limits of such tools? What are the ethics of allowing AI tools to suggest different courses of action, or aid in decision-making? And does AI offer too much promise for individuals? We draw as to how AI software and technology might best be developed in the future.

Coeckelbergh, M. “AI for climate: freedom, justice, and other ethical and political challenges” AI Ethics (2020). https://doi.org/10.1007/s43681-020-00007-2

Artificial intelligence can and should help to build a greener, more sustainable world and to deal with climate change, but these opportunities also raise ethical and political issues that need to be addressed. This article discusses these issues, with a focus on problems concerning freedom and justice at a global level, and calls for responsible use of AI for climate in the light of these challenges.

Hickok, M. “Lessons learned from AI ethics principles for future actions” AI Ethics (2020). https://doi.org/10.1007/s43681-020-00008-1

The use of AI systems is significantly more prevalent in recent years, and the concerns on how these systems collect, use and process big data has also increased. To address these concerns and advocate for ethical and responsible AI development and implementation, NGOs, research centers, private companies, and governmental agencies have published more than 100 AI ethics principles and guidelines. Lessons must be learned from the shortcomings of AI ethics principles to ensure that future investments, collaborations, standards, codes, and legislation reflect the diversity of voices and incorporate the experiences of those who are already impacted by AI.

Fall Nuggets

USTPC Panel on Section 230

On November 18 from 5:00 to 6:30 PM EST, experts from ACM’s US Technology Policy Committee (USTPC) will discuss the legal liability of Internet platforms such as Facebook and Twitter under Section 230 of the Communications Decency Act. The USTPC panelists are Andy Grosso (Moderator), Mark Rasch, Pam Samuelson, Richard M. Sherman, and Danny Weitzner.

Biden and Science

Participants in a Science and Business Webcast urged that a global assembly “should press leaders of the big industrial nations to open – or re-open – their research systems, while also ensuring that COVID-19 vaccines are freely available to everyone in the world. An international summit.” About an international summit, Robert-Jan Smits, former director-general of the European Commission’s research and innovation directorate said it, “would really show that senior leaders are turning the page,”

Center for Data Innovation On the EU Data Governance Act

“The European Commission is planning to release its Data Governance Act to facilitate data sharing within the EU. The goal is to increase data sharing among businesses, make more public-sector data available for reuse, and foster data sharing of personal data, including for ‘altruistic’ purposes. While the goals of the act are commendable, many of the specific policies outlined in a draft would create a new data localization requirement, undermine the EU’s commitments to digital free trade, and contradict its open data principles.”

AI Data

Confusion in the popular media about terms such as algorithm and what constitutes AI technology cause critical misunderstandings among the public and policymakers. More importantly, the role of data is often ignored in ethical and operational considerations. Even if AI systems are perfectly built, low quality and biased data cause unintentional and even intentional hazards.

Language Models and Data

A generative pre-trained transformer GPT-3 is currently in the news. For example, James Vincent in the July 30, 2020, article in The Verge writes about GPT-3, which was created by OpenAI. Language models, GPT-3 the current ultimate product, have ethics issues on steroids for products being made. Inputs to the system have all the liabilities discussed about Machine Learning and Artificial Neural Network products. The dangers of bias and mistakes are raised in some writings but are likely not a focus among the wide range of enthusiastic product developers using the open-source GPT-3. Language models suggest output sequences of words given an input sequence. Thus, samples of text from social media can be used to produce new text in the same style as the author and potentially can be used to influence public opinion. Cases have been found of promulgating incorrect grammar and misuse of terms based on poor quality inputs to language models. An article by David Pereira includes examples and comments on the use of GPT-3. The article “GPT-3: an AI Game-Changer or an Environmental Disaster?” by John Naughton gives examples of and commentary on results from GPT-3.

Data Governance

A possible meta solution for policymakers to keep up with technological advances is discussed by Alex Woodie in “AI Ethics and Data Governance: A Virtuous Cycle.”

He quotes James Cotton, who is the international director of the Data Management Centre of Excellence at Information Builders’ Amsterdam office: “as powerful as the AI technology is, it can’t be implemented in an ethical manner if the underlying data is poorly managed and badly governed. It’s critical to understand the relationship between data governance and AI ethics. One is foundational for the other. You can’t preach being ethical or using data in an ethical way if you don’t know what you have, where it came from, how it’s being used, or what it’s being used for.”

USTPC in the News

Overview

The ACM’s US Technology Policy Committee (USTPC) has been very active in July already! The contributions and visibility of USTPC as a group and as individual members are very welcome and impressive. The following list has links to highly-recommended reading.

Amicus Brief: USTPC Urges Narrower Definition of Computer Fraud and Abuse Act

ACM’s USTPC filed an amicus curiae (“friend of the court”) brief with the United States Supreme Court in the landmark case of Van Buren v. United States. “Van Buren marks the first time that the US Supreme Court has reviewed the Computer Fraud and Abuse Act (CFAA), a 1986 law that was originally intended to punish hacking. In recent years, however, the CFAA has been used to criminally prosecute both those who access a computer system without permission, as well as those who have permission but exceed their authority to use a database once logged in.”

USTPC Statement on Face Recognition

(USTPC) has assessed the present state of facial recognition (FR) technology as applied by government and the private sector. The Committee concludes that, “when rigorously evaluated, the technology too often produces results demonstrating clear bias based on ethnic, racial, gender, and other human characteristics recognizable by computer systems. The consequences of such bias, USTPC notes, frequently can and do extend well beyond inconvenience to profound injury, particularly to the lives, livelihoods and fundamental rights of individuals in specific demographic groups, including some of the most vulnerable populations in our society.”
See the NBC news article.

Barbara Simons recipient of the 2019 ACM Policy Award

USTPC’s Barbara Simons, founder of USTPC predecessor USACM, is the recipient of the 2019 ACM Policy Award for “long-standing, high-impact leadership as ACM President and founding Chair of ACM’s US Public Policy Committee (USACM), while making influential contributions to improve the reliability of and public confidence in election technology. Over several decades, Simons has advanced technology policy by founding and leading organizations, authoring influential publications, and effecting change through lobbying and public education.”
Congratulations, Barbara!

Potential New Issues

ACM Urges Preservation of Temporary Visa Exemptions for Nonimmigrant Students. Harvard filing is a complaint for declaratory and injunctive relief.

This issue may have dramatic impacts on university research and teaching this fall.

Thank you USTPC for your hard work and representation of ACM to policymakers!

AI and Facial Recognition

AI in Congress

Politico reports on two separate bills introduced Thursday, June 2. (See the section entitled “Artificial Intelligence: Let’s Do the Thing”.)

The National AI Research Resource Task Force Act. “The bipartisan, bicameral bill introduced by Reps. Anna Eshoo, (D-Calif.), Anthony Gonzalez (R-Ohio), and Mikie Sherrill (D-N.J.), along with companion legislation by Sens. Rob Portman (R-Ohio) and Martin Heinrich(D-N.M.), would form a committee to figure out how to launch and best use a national AI research cloud. Public and private researchers and developers from across the country would share this cloud to combine their data, computing power and other resources on AI. The panel would include experts from government, academia and the private sector.”

The Advancing Artificial Intelligence Research Act. “The bipartisan bill introduced by Senate Commerce Chairman Roger Wicker (R-Miss.), Sen. Cory Gardner (R-Colo.) and Gary Peters (D-Mich.), a founding member of the Senate AI Caucus, would create a program to accelerate research and development of guidance around AI at the National Institute of Standards and Technology. It would also create at least a half-dozen AI research institutes to examine the benefits and challenges of the emerging technology and how it can be deployed; provide funding to universities and nonprofits researching AI; and launch a pilot at the National Science Foundation for AI research grants.”

Concerns About Facial Recognition (FR): Discrimination, Privacy, and Democratic Freedom

While including ethical and moral issues, a broader list of issues is concerning to citizens and policymakers about face recognition technology and AI. Areas of concerns include accuracy; surveillance; data storage, permissions, and access; discrimination, fairness, and bias; privacy and video recording without consent; democratic freedoms, including right to choose, gather, and speak; and abuse of technology such as non-intended uses, hacking, and deep fakes. Used responsibly and ethically, face recognition can be valuable for finding missing people, responsible policing and law enforcement, medical uses, healthcare, virus tracking, legal system and court uses, and advertising. Various guidelines by organizations such as the AMA and legislation like S.3284 – Ethical Use of Facial Recognition Act are being developed to encourage the proper use of AI and face recognition.

Some of the above issues do specifically require ethical analysis as in the following by Yaroslav Kuflinski:

Accuracy — FR systems naturally discriminate against non-whites, women, and children, presenting errors of up to 35% for non-white women.

Surveillance issues — concerns about “big brother” watching society.

Data storage — use of images for future purposes stored alongside genuine criminals.

Finding missing people — breaches of the right to a private life.

Advertising — invasion of privacy by displaying information and preferences that a buyer would prefer to keep secret.

Studies of commercial systems are increasingly available, for example an analysis of Amazon Rekognition.

Biases deriving from sources of unfairness and discrimination in machine learning have been identified in two areas: the data and the algorithms.  Biases in data skew what is learned in machine learning methods, and flaws in algorithms can lead to unfair decisions even when the data is unbiased. Intentional or unintentional biases can exist in the data used to train FR systems.

New human-centered design approaches seek to provide intentional system development steps and processes in collecting data and creating high quality databases, including the elimination of naturally occurring bias reflected in data about real people.

Bias That Pertains Especially to Facial Recognition (Mehrabi, et al. and Barocas, et al.)

Direct Discrimination: “Direct discrimination happens when protected attributes of individuals explicitly result in non-favorable outcomes toward them”.  Some traits like race, color, national origin, religion, sex, family status, disability, exercised rights under CCPA , marital status, receipt of public assistance, and age are identified as sensitive attributes or protected attributes in the machine learning world.                       

Indirect Discrimination: Even if sensitive or protected attributes are not used against an individual, indirect discrimination can still happen. For example, residential zip code is not categorized as a protected attribute, but from the zip code one might infer race, which is a protected attribute. So, “protected groups or individuals still can get treated unjustly as a result of implicit effects from their protected attributes”.

Systemic Discrimination: “policies, customs, or behaviors that are a part of the culture or structure of an organization that may perpetuate discrimination against certain subgroups of the population”.

Statistical Discrimination: In law enforcement, racial profiling is an example of statistical discrimination. In this case, minority drivers are pulled over more than compared to white drivers — “statistical discrimination is a phenomenon where decision-makers use average group statistics to judge an individual belonging to that group.”

Explainable Discrimination: In some cases, discrimination can be explained using attributes like working hours and education, which is legal and acceptable. In “the UCI Adult dataset [6], a widely-used dataset in the fairness domain, males on average have a higher annual income than females; however, this is because on average females work fewer hours than males per week. Work hours per week is an attribute that can be used to explain low income. If we make decisions without considering working hours such that males and females end up averaging the same income, we could lead to reverse discrimination since we would cause male employees to get lower salary than females.                             

Unexplainable Discrimination: This type of discrimination is not legal as explainable discrimination because “the discrimination toward a group is unjustified”.

How to Discuss Facial Recognition

Recent controversies about FR mix technology issues with ethical imperatives and ignore that people can disagree on which are the “correct” ethical principles. A recent ACM tweet on FR and face masks was interpreted in different ways and ACM issued an official clarification. A question that emerges is if AI and other technologies should be, and can be, banned rather than controlled and regulated.

In early June, 2020, IBM CEO Arvind Krishna said in a letter to Congress that IBM is exiting the facial recognition business and asking for reforms to combat racism: “IBM no longer offers general purpose IBM facial recognition or analysis software. IBM firmly opposes and will not condone uses of any technology, including facial recognition technology offered by other vendors, for mass surveillance, racial profiling, violations of basic human rights and freedoms, or any purpose which is not consistent with our values and Principles of Trust and Transparency,” Krishna said in his letter to members of congress, “We believe now is the time to begin a national dialogue on whether and how facial recognition technology should be employed by domestic law enforcement agencies.”

The guest co-author of this series of blog posts on AI and bias is Farhana Faruqe, doctoral student in the George Washington University Human-Technology Collaboration program.

Policy and AI Ethics

The Alan Turing Institute Public Policy Programme

Among the complexities of public policy making, the new world of AI and data science requires careful consideration of ethics and safety in addressing complex and far-reaching challenges in the public domain. Data and AI systems lead to opportunities that can produce both good and bad outcomes. Ethical and safe systems require intentional processes and designs for organizations responsible for providing public services and creating public policies. An increasing amount of research focuses on developing comprehensive guidelines and techniques for industry and government groups to make sure they consider the range of issues in AI ethics and safety in their work. An excellent example is the Public Policy Programme at The Alan Turing Institute under the direction of Dr. David Leslie [1]. Their work complements and supplements the Data Ethics Framework [2], which is a practical tool for use in any project initiation phase. Data Ethics and AI Ethics regularly overlap.

The Public Policy Programme describes AI Ethics as “a set of values, principles, and techniques that employ widely accepted standards of right and wrong to guide moral conduct in the development and use of AI technologies. These values, principles, and techniques are intended both to motivate morally acceptable practices and to prescribe the basic duties and obligations necessary to produce ethical, fair, and safe AI applications. The field of AI ethics has largely emerged as a response to the range of individual and societal harms that the misuse, abuse, poor design, or negative unintended consequences of AI systems may cause.”

They cite the following as some of the most consequential potential harms:

  • Bias and Discrimination
  • Denial of Individual Autonomy, Recourse, and Rights
  • Non-transparent, Unexplainable, or Unjustifiable Outcomes
  • Invasions of Privacy
  • Isolation and Disintegration of Social Connection
  • Unreliable, Unsafe, or Poor-Quality Outcomes

The Ethical Platform for the Responsible Delivery of an AI Project, strives to enable the “ethical design and deployment of AI systems using a multidisciplinary team effort. It demands the active cooperation of all team members both in maintaining a deeply ingrained culture of responsibility and in executing a governance architecture that adopts ethically sound practices at every point in the innovation and implementation lifecycle.” The goal is to “unite an in-built culture of responsible innovation with a governance architecture that brings the values and principles of ethical, fair, and safe AI to life.”

[1] Leslie, D. (2019). Understanding artificial intelligence ethics and safety: A guide for the responsible design and implementation of AI systems in the public sector. The Alan Turing Institute. https://doi.org/10.5281/zenodo.3240529

[2] Data Ethics Framework (2018). https://www.gov.uk/government/publications/data-ethics-framework/data-ethics-framework.

Principled Artificial Intelligence

In January, 2020, the Berkman Klein Center released a report by Jessica Fjeld and Adam Nagy “Mapping Consensus in Ethical and Rights-Based Approaches to Principles for AI”, which summarizes contents of 36 documents on AI principles.

This work acknowledges the surge in frameworks based on ethical and human rights to guide the development and use of AI technologies.  The authors focus on understanding ethics efforts in terms of eight key thematic trends:  

  • Privacy
  • Accountability
  • Safety & security
  • Transparency & explainability
  • Fairness & non-discrimination
  • Human control of technology
  • Professional responsibility
  • Promotion of human values

They report “our analysis examined the forty-seven individual principles that make up the themes, detailing notable similarities and differences in interpretation found across the documents. In sharing these observations, it is our hope that policymakers, advocates, scholars, and others working to maximize the benefits and minimize the harms of AI will be better positioned to build on existing efforts and to push the fractured, global conversation on the future of AI toward consensus.”

Human-Centered AI

Prof. Ben Shneiderman recently presented his extensive work “Human-Centered AI: Trusted, Reliable & Safe” at the University of Arizona’s NSF Workshop on “Assured Autonomy”.  His research emphasizes human autonomy as opposed to the popular notion of autonomous machines. His Open Access paper quickly drew 3200+ downloads. The ideas are now available in the International Journal of Human–Computer Interaction. The abstract is as follows: “Well-designed technologies that offer high levels of human control and high levels of computer automation can increase human performance, leading to wider adoption. The Human-Centered Artificial Intelligence (HCAI) framework clarifies how to (1) design for high levels of human control and high levels of computer automation so as to increase human performance, (2) understand the situations in which full human control or full computer control are necessary, and (3) avoid the dangers of excessive human control or excessive computer control. The methods of HCAI are more likely to produce designs that are Reliable, Safe & Trustworthy (RST). Achieving these goals will dramatically increase human performance, while supporting human self-efficacy, mastery, creativity, and responsibility.”

COVID AI

AI is in the news and in policy discussions regarding COVID-19, both about ways to help fight the pandemic and in terms of ethical issues that policymakers should address. Michael Corkery and David Gelles in the NY Times article “Robots Welcome to Take Over, as Pandemic Accelerates Automation”, suggest that “social-distancing directives, which are likely to continue in some form after the crisis subsides, could prompt more industries to accelerate their use of automation.” An MIT Technology Review article by Genevieve Bell, “We need mass surveillance to fight covid-19—but it doesn’t have to be creepy” looks at the pros and cons of AI technology and if we now have the chance to “reinvent the way we collect and share personal data while protecting individual privacy.”

Public Health and Privacy Issues

Liza Lin and Timothy W. Martin in “How Coronavirus Is Eroding Privacy” write about how technology is being developed to track and monitor individuals for slowing the pandemic, but that this “raises concerns about government overreach.” Here is an excerpt from that WSJ article: “Governments worldwide are using digital surveillance technologies to track the spread of the coronavirus pandemic, raising concerns about the erosion of privacy. Many Asian governments are tracking people through their cellphones to identify those suspected of being infected with COVID-19, without prior consent. European countries are tracking citizens’ movements via telecommunications data that they claim conceals individuals’ identities; American officials are drawing cellphone location data from mobile advertising firms to monitor crowds, but not individuals. The biggest privacy debate concerns involuntary use of smartphones and other digital data to identify everyone with whom the infected had recent contact, then testing and quarantining at-risk individuals to halt the further spread of the disease. Public health officials say surveillance will be necessary in the months ahead, as quarantines are relaxed and the virus remains a threat while a vaccine is developed.

“In South Korea, investigators scan smartphone data to find within 10 minutes people who might have caught the coronavirus from someone they met. Israel has tapped its Shin Bet intelligence unit, usually focused on terrorism, to track down potential coronavirus patients through telecom data. One U.K. police force uses drones to monitor public areas, shaming residents who go out for a stroll.

“The Covid-19 pandemic is ushering in a new era of digital surveillance and rewiring the world’s sensibilities about data privacy. Governments are imposing new digital surveillance tools to track and monitor individuals. Many citizens have welcomed tracking technology intended to bolster defenses against the novel coronavirus. Yet some privacy advocates are wary, concerned that governments might not be inclined to unwind such practices after the health emergency has passed.

“Authorities in Asia, where the virus first emerged, have led the way. Many governments didn’t seek permission from individuals before tracking their cellphones to identify suspected coronavirus patients. South Korea, China and Taiwan, after initial outbreaks, chalked up early successes in flattening infection curves to their use of tracking programs.

“In Europe and the U.S., where privacy laws and expectations are more stringent, governments and companies are taking different approaches. European nations monitor citizen movement by tapping telecommunications data that they say conceals individuals’ identities.

American officials are drawing cellphone location data from mobile advertising firms to track the presence of crowds—but not individuals. Apple Inc. and Alphabet Inc.’s Google recently announced plans to launch a voluntary app that health officials can use to reverse-engineer sickened patients’ recent whereabouts—provided they agree to provide such information.”

Germany Changes Course on Contact Tracing App

Politico reports that “the German government announced today” (4/26) “that Berlin would adopt a ‘decentralized’ approach to a coronavirus contact-tracing app — now backing an approach championed by U.S. tech giants Apple and Google. ‘We will promote the use of a consistently decentralized software architecture for use in Germany,’ the country’s Federal Health Minister Jens Spahn said on Twitter, echoing an interview in the Welt am Sonntag newspaper. Earlier this month, Google and Apple announced they would team up to unlock their smartphones’ Bluetooth capabilities to allow developers to build interoperable contact tracing apps. Germany is now abandoning a centralized approach spearheaded by the German-led Pan-European Privacy-Preserving Proximity Tracing (PEPP-PT) project. Berlin’s U-turn comes after a group of six organizations on Friday urged Angela Merkel’s government to reassess plans for a smartphone app that traces potential coronavirus infections, warning that it does not do enough to protect user data.”

NSF Program on Fairness in Artificial Intelligence (FAI) in Collaboration with Amazon

A new National Science Foundation solicitation NSF 20-566 has been announced by the Directorate for Computer and Information Science and Engineering, Division of Information and Intelligent Systems, Directorate for Social, Behavioral and Economic Sciences, and Division of Behavioral and Cognitive Sciences.

Bias and Fairness

Today’s post has AI and Policy news updates and our next installment on Bias and Policy: the fairness component.

News Items for February, 2020

  • OECD launched the OECD.AI Observatory, an online platform to shape and share AI policies across the globe. 
  • The White House released the American Artificial Intelligence Initiative:Year One Annual Report and supported the OECD policy.

Bias and Fairness

In terms of decision-making and policy, fairness can be defined as “the absence of any prejudice or favoritism towards an individual or a group based on their inherent or acquired characteristics”.  Six of the most used definitions are equalized odds, equal opportunity, demographic parity, fairness through unawareness or group unaware, treatment equality. 

The concept of equalized odds and equal opportunity is that individuals who qualify for a desirable outcome should have an equal chance of being correctly assigned regardless of an individual’s belonging to a protected or unprotected group (e.g., female/male). The additional concepts “demographic parity” and “group unaware” are illustrated by the Google visualization research team with nice visualizations using an example “simulating loan decisions for different groups”. The focus of equal opportunity is on the outcome of the true positive rate of the group.

On the other hand, the focus of the demographic parity is on the positive rate only. Consider a loan approval process for two groups: group A and group B. For demographic parity, the overall number of approved loans should be equal in both group A and group B regardless of a person belonging to a protected group. Since the focus for demographic parity is on overall loan approval rate, the rate should be equal for both the groups. Some people in group A who would pay back the loan might be disadvantaged compared to the people in group B who might not pay back the loan.  However, the people in group A will not be at a disadvantage in the equal opportunity concept, since this concept focuses on true positive rate. As an example of fairness through unawareness “an algorithm is fair as long as any protected attributes A are not explicitly used in the decision-making process”.

All of the fairness concepts or definitions either fall under individual fairness, subgroup fairness or group fairness. For example, demographic parity, equalized odds, and equal opportunity are the group fairness type; fairness through awareness falls under the individual type where the focus is not on the overall group.

A definition of bias can be in three categories: data, algorithmic, and user interaction feedback loop:
Data — behavioral bias, presentation bias, linking bias, and content production bias;
Algoritmic — historical bias, aggregation bias, temporal bias, and social bias falls
User Interaction — popularity bias, ranking bias, evaluation bias, and emergent bias.

Bias is a large domain with much to explore and take into consideration. Bias and public policy will be further discussed in future blog posts.

This series of posts on Bias has been co-authored by Farhana Faruqe, doctoral student in the GWU Human-Technology Collaboration group.

References 
 [1] Ninareh Mehrabi, Fred Morstatter, Nripsuta Saxena, Kristina Lerman, and Aram Galstyan. A survey on bias and fairness in machine learning. CoRR, abs/1908.09635, 2019.
[2] Moritz Hardt, Eric Price, , and Nati Srebro. 2016. Equality of Opportunity in Supervised Learning. In Advances in Neural Information Processing Systems 29, D. D. Lee, M. Sugiyama, U. V. Luxburg, I. Guyon, and R. Garnett (Eds.). Curran Associates, Inc., 3315–3323. http://papers.nips.cc/paper/ 6374-equality-of-opportunity-in-supervised-learning.pdf
[3] Martin Wattenberg, Fernanda Viegas, and Moritz Hardt. Attacking discrimination with smarter machine learning. Accessed at https://research.google.com/bigpicture/attacking-discrimination-in-ml/, 2016

Discrimination and Bias

Our current public policy posts, focused on ethics and bias in current and emerging areas of AI, build on the work “A Survey on Bias and Fairness in Machine Learning” by Ninareh Mehrabi, et al. and resources provided by Barocas, et al. The guest co-author of this series of blog posts on AI and bias is Farhana Faruqe, doctoral student in the George Washington University Human-Technology Collaboration program. We look forward to your comments and suggestions.

Discrimination, unfairness, and bias are terms used frequently these days in the context of AI and data science applications that make decisions in the everyday lives of individuals and groups. Machine learning applications depend on datasets that are usually a reflection of our real world in which individuals have intentional and unintentional biases that may cause discrimination and unfair actions. Broadly, fairness is the absence of any prejudice or favoritism towards an individual or a group based on their intrinsic or acquired traits in the context of decision-making.

Today’s blog post focuses on discrimination, which Ninareh Mehrabi, et al. describe as follows:

Direct Discrimination: “Direct discrimination happens when protected attributes of individuals explicitly result in non-favorable outcomes toward them.”  Some traits like race, color, national origin, religion, sex, family status, disability, marital status, recipient of public assistance, and age are identified as sensitive attributes or protected attributes in the machine learning world.  It is not legal to discriminate against these sensitive attributes, which are listed by the FHA and Equal Credit Opportunity Act (ECOA).                

Indirect Discrimination: Even if sensitive or protected attributes are not used against an individual, still indirect discrimination can happen. For example, residential zip code is not categorized as a protected attribute, but from the zip code one may find out about race which is a protected attribute. So, “protected groups or individuals still can get treated unjustly as a result of implicit effects from their protected attributes.”

Systemic Discrimination. In the nursing profession, the custom is to expect a nurse to be a woman. So, excluding qualified male nurses for nursing position is an example of systematic discrimination. Systematic discrimination is defined as “policies, customs, or behaviors that are a part of the culture or structure of an organization that may perpetuate discrimination against certain subgroups of the population”.                                                                                                                              
Statistical Discrimination: In law enforcement, racial profiling is an example of statistical discrimination. In this case, minority drivers are pulled over more often than white drivers. The authors define “statistical discrimination is a phenomenon where decision-makers use average group statistics to judge an individual belonging to that group.”

Explainable Discrimination: In some cases, “discrimination can be explained using attributes” like working hours and education, which is legal and acceptable as well. In a widely used dataset in the fairness domain, males on average have a higher annual income than females because on average females work fewer hours per week than males do. Decisions made without considering working hours could lead to discrimination.                     

Unexplainable Discrimination: This type of discrimination is not legal as explainable discrimination because “the discrimination toward a group is unjustified”. Some researchers have introduced techniques during data preprocessing and training to remove unexplainable discrimination.   

To understand bias in techniques such as machine learning, we will discuss in our next blog post another important aspect: fairness.