Social Work, Artificial Intelligence and Child Welfare – an Ongoing Scan of the Literature

Last week, while at the Exponential Medicine conference (full download to follow soon), I heard about the involvement of Artificial Intelligence (AI) in U.S. child welfare practice. This was new to me – I wondered why I hadn’t heard more about it.

I immediately went to a colleague who I knew was very active in this space, Dr. Melanie Sage, and asked her if she knew about use of AI in this way. She had a great many resources, ideas and connections in this area. Together we put together a resource list and annotated bibliography for scholars, teachers, community members, and students might be interested in (Melanie’s contributions are based on her scholarly work in this area, mine are based on never-ending curiosity and my futures activity). We both believe we’d like to see more social workers getting curious, getting creative and applying their knowledge, values and skills as social workers to make sure these approaches are used with the highest ethical dimension possible. In truth, there is no indication that artificial intelligence is going away – and all signs point to greater expansion with these tools. We are literally watching the path being built while we walk on it. We both agreed we’d like that building process to be one informed by the literature – so here’s a place to start.

Where and how can tools like artificial intelligence be used to support positive outcomes for families and children in the child welfare system? What are the risks and potential complications that could interfere with our good intentions? We are just beginning to find out. There are opportunities to truly revolutionize outcomes for the better – developing analytic capabilities we have never had before using information in all new ways to improve our practice. And there are risks to use of these tools that are also covered in the literature and well-documented. Come learn with us and let’s use technology to make a better world.

Want an overview of the fundamentals of AI? Check out a previous blog post covering the basics!

Reflections and Take-Away from Amy Webb’s Book “The Big Nine”

“The future we all want to live in won’t just show up, fully formed. We need to be courageous. We must take responsibility for our actions.” – Amy Webb

I recently finished reading Amy Webb’s fine book “The big nine: How the tech titans thinking machines could warp humanity,” (2019). New York: Public Affairs/Hachette Book Group.

The book is a history and evolution of the power of “the big nine” tech companies (6 US-based, 3 in China), with a primary focus on the power and possibilities of artificial intelligence. It takes a deep look at all the power, opportunities, possibilities (both positive and devastating) that AI brings now and into the future.

The Big Nine defies a “simple” framework (AI is good/AI is bad). Rather it focuses on the idea that AI is almost incomprehensibly powerful and requires the responsible attention of individuals, communities and governments to assure that the highest ideals and possibilities are achieved and the greatest threats are reduced/eliminated (almost as if one might think of the way that we think about power/possibility of nuclear power – though the developmental trajectories have distinct differences).

From a social work perspective, the focus intersects with our own thinking/imagining of the “future” of social justice, human well-being and equity work. What is a future in which a few powerful (largely white, male, economically dominant and western in the US) construct underlying structures and digital machinery that decides, sorts, and controls much of the workings of modern life? How might existing inequities be replicated, multiplied – or conversely, interrupted and resolved? These are essential concerns that social work would be well-advised to factor into the way we think about the future and our work in it. How will these mechanisms (or have already begun to) (re) arrange modern life, who will continue to win and lose, and how will those trajectories play out according to the way social work thinks about ourselves and the work we aspire to do? Likely these will be a combination of ways that we are professionally familiar with (poverty, structural violence, “isms” and the like) as well as new types or variations of oppressions that we can only begin to predict and understand. My recent blog post on algorithmic transparency, bias and justice goes into some of these issues in more detail.

Our values, knowledge and skills regarding the importance and processes of engaging community voices, interrupting oppression, building more just and liberatory structures, and recognizing and addressing structural barriers to well-being could all be important skills as increasing pressure builds regarding recognizing the human rights issues associated with growth in tech that is not reflective of the well-being of all. But we will need to be intentional regarding our needed learning curve to remain relevant in these complex new spaces. I found this book to advance my own thinking/understanding regarding how vast and complex discussions of “big tech” and AI can be, and yet largely understandable using our own frames of political economy, human rights and social work ethics – just in new spaces and new ways. Social workers belong in this conversation, and this blog remains a call to action and invitation to continue dialogue about how we might best do that.

While the book as a whole is readable including three primary sections: 1) overview and evolution of tech in the modern world (formidably challenge for the non-tech reader but she does a fine job of keeping it accessible), 2) a fascinating, inspiring and sobering deep presentation of three possible “futures” concerning AI. These scenarios are crafted with the intention of fully exploring various possibilities that exist for humanity based on decisions that are made (as Ms. Webb might say) while our ability to do so is still collectively within our grasp and 3) a final section that lays out an action plan and analysis of what needs to be done to optimize all that AI has to offer, while simultaneously building a new set of global policy guardrails to protect us, in some respects, from ourselves and the worst of the risks that are increasingly apparent in the rapid evolution of these technologies. The purpose of this post is to share what I considered to be the most substantive part of the book, which is Ms. Webb’s suggestion that to succeed in the years ahead with the complexities (and risks) that AI introduces into our world, an international body comprised of tech leaders and “AI researchers, sociologists, economists, game theorists, futurists, political scientists” (p. 237) along with government leaders, and that these members reflect the “socioeconomic, gender, race, religious, political and sexual diversity” of the world (p. 237).

She calls this governing/regulatory body the Global Alliance on Intelligence Augmentation (GAIA), and their core aspirational purpose would be to collectively “facilitate and cooperate on share AI initiatives and policies” (p. 237) and to affirm and create structures to consider, operationalize and protect AI as a public good. In essence, she suggests that these tools are rapidly becoming too powerful to be left merely to the devices of private, corporate and market forces.

Here is an excerpt that clarifies what I consider to be the most important elements of this effort she proposes – which in itself is a fascinating “thought experiment” about what might be to come. I hope that we move towards this kind of global dialogue sooner rather than later – and I hope that we as social workers – can find ourselves as helpful, informative, relevant change agents, social scientists, and supporters of human well-being in an increasingly complicated world.

“GAIA should be considered a framework of rights that balances individual liberties with the greater, global good. It would be better to establish a framework that’s strong on ideals but can be more flexible in interpretation as AI matures. Member organizations would have to demonstrate they are in compliance or face being removed from GAIA. Any framework should include the following principles:

  1. Humanity should always be at the center of AI’s development.
  2. AI systems should be safe and secure. We should be able to independently verify their safety and security.
  3. The Big Nine – including its investors, employees, and the governments it works within – must prioritize safety above speed. Any team working on an AI system – even those outside the Big Nine – must not cut corners in favor of speed. Safety must be demonstrated and discernable by independent observers.
  4. If an AI system causes harm, it should be able to report out what went wrong, and there should be a governance process in place to discuss and mitigate damage.
  5. AI should be explainable. Systems should carry something akin to a nutritional label, detailing the training data used, the processes used for learning, the real-world data being used in applications and the expected outcomes. For sensitivity or proprietary systems, trusted third parties should be able to assess and verify an AI’s transparency.
  6. Everyone in the AI ecosystem – Big Nine employees, managers, leaders, board members; startups (entrepreneurs and accelerators); investors (venture capitalists, private equity firms, institutional investors, and individual shareholders); teachers and graduate students; and anyone else working in AI – must recognize that they are making ethical decisions all the time. They should be prepared to explain all of the decisions they’ve made during the development, testing and deployment processes.
  7. The Human Values Atlas* should be adhered to for all AI projects. Even narrow AI applications should demonstrate that the atlas has been incorporated.
  8. There should be a published, easy-to-find code of conduct governing all people who work on AI and its design, build and deployment. The code of conduct should also govern investors.
  9. All people should have the right to interrogate AI systems. What an AI’s true purpose is, what data it uses, how it reaches its conclusions, and who sees results should be made fully transparent in a standardized format.
  10. The terms of service for an AI application-or any service that uses AI – should be written in language plain enough that a third grader can comprehend it. It should be available in every language as soon as the application goes live.
  11. PDR’s (personal data records) should be opt-in and developed using a standard format, they should be interoperable, and individual people should retain full ownership and permission rights. Should PDR’s become heritable, individual people should be able to decide the permissions and uses of their data.
  12. PDR’s should be decentralized as much as possible, ensuring that no one party has complete control. The technical group that designs our PDRs should include legal and nonlegal experts alike: whitehat (good) hackers, civil rights leaders, government agents, independent data fiduciaries, ethicists, and other professionals working outside of the Big Nine.
  13. To the extent possible, PDRs should be protected against enabling authoritarian regimes.
  14. There must be a system of public accountability and an easy method for people to receive answers to questions about their data and how it is mined, refined and use throughout AI systems.
  15. All data should be treated fairly and equally, regardless of nationality, race, religion, sexual identity, gender, political affirmations, or other uniques beliefs,” (pp. 240-242).

*The idea of a “human values atlas” is presented earlier in the book as the formidable and complex but essential task of creating a living and shared communication/document about what is most centrally valued by humans across cultures and nationalities. This atlas would guide much of the future work in the AI space – without it – we are as Ms. Webb suggests, ceding authority for these matters to potentially conflicting and hidden/opaque corporate forces. She discusses this in greater detail on pages 239-240 of the book.

Here is a 15-minute interview with Ms. Webb on a recent PBS spot.

For the reader’s convenience, here are a couple of additional reviews of this book:

Technology Review

Fast Company

Venture Beat

Wired

Finally here is some information about recent and current U.S. federal activity on this issue:

Will Trump’s new artificial intelligence initiative make the U.S. the world leader in AI? (2019)

President Obama’s artificial intelligence, automation and the economy plan (2016)

Algorithmic Transparency, Bias and Justice

Algorithms are a huge part of modern life. So much so that we sometimes forget they have arrived. Indeed they are primarily “invisible” to everyday people, working behind the scenes to sort data and make decisions that reflect the opinions of a few algorithm designers behind the scenes. Sometimes these algorithms can be life changing/life saving, for example when cancer diagnosis can be made through a combination of machine learning and algorithms that can scan hundreds of thousands of xrays to detect the tiniest irregularity that a human might miss. But other times, like racially biased facial recognition software that might inaccurately identify someone as a criminal suspect – are much more concerning. Increasingly, the ideas of “algorithmic transparency,” “algorithmic racism/bias,” and “algorithmic justice” have come into more prevalent conversation among social justice circles.

There is much learning and development going on with regard to this topic. Of all the “future facing” topics one might consider in terms of urgent need for attention in social work – in my estimation – this is one of the most important. As the rate of adoption of new technologies (most often emerging from the private sector) continues to accelerate, algorithms that don’t incorporate ethical and bias-free dimensions are a frequent point of discussion among social justice advocates. What is the pathway forward and how do we continue to increase social work practice and research attention in this area?

I would suggest that this is the most under-discussed ethical challenge of the future for the profession of social work. We need to dramatically increase the depth, range and focus of our ethical evolution to participate in and shape the future of these technologies that work for people and that prevent harm and injustice. We should concern ourselves with identifying how and where algorithms are starting to emerge and be active in our social work practice spaces (clinical and macro). Collectively – we are starting to develop a shared and critical literacy regarding these important and ubiquitous forces, and challenge a need for clear and explicit ethical guidelines/rules.

For those who are completely new to this topic, here’s a great primer.

While there are pockets of enthusiasm for dialogue about these developments in social work, we have a long way to go to assert where and how we can operate most ethically – and what that looks like given the changing dynamics at play.

Here’s a reading/resource list of resources to get started – with great respect for the groundbreaking work of all who have been leaders in this space.

  • Dr. Desmond Patton is an Associate Professor of Social Work at Columbia University in New York City. I’ve previously listed his work on my blog but want to underscore the significant leadership he’s contributed within social work to this topic. Here’s a recent article he put together for Medium. He’s also the Principal Investigator of the Safe Lab project at Columbia which is a research initiative focused on examining the ways in which youth of color navigate violence on and offline.
  • Data for Black Lives is a national network of over 4,000 activists, organizers, and scientists using data science to create concrete and measurable change in the lives of Black people. For far too long, data has been weaponized against Black communities – from redlining to predictive policing, credit scoring and facial recognition. But we are charting out a new era, where data is a tool for profound social change. (From their website here!)
  • The Institute for the Future has developed an “Ethical OS” toolkit to provide a structure for tech experts to use to deepen their adherence to ethical principles while developing tech tools. Check it out here.

These are the books currently on my shelf on this topic:

Eubanks, V. (2018). Automating inequality: How high tech tools profile, punish and police the poor. New York: St. Martin’s Press. Review here.

Lane, J. (2019). The digital street. New York: Oxford Press. Review here.

Noble, S.U. (2018). Algorithms of oppression: How search engines reinforce racism. New York: New York University Press. Review here.

O’Neill, C. (2016). Weapons of math destruction: How big data increases inequality and threatens democracy. New York: Broadway Books. Review here – scroll down where her TED talk is included.

Also, I’ve collected numerous recent articles about bias, “isms” and ethics concerns regarding algorithmic transparency/bias as follows:

Behind every robot is a human (2019)

The new digital divide is between people who opt out of algorithms and those who don’t (2019)

Collection of new articles from the Brookings Institute regarding AI and the future (2019)

Artificial intelligence is ripe for abuse, tech researcher warns: A fascist’s dream (2019)

Algorithmic Accountability Act (2019)

Amazon Alexa launches its first HIPAA compliant medical unit (2019)

Facial recognition is big tech’s latest toxic gateway app (2019)

That mental health app might share your data without telling you (2019)

Europe is making AI rules now to avoid a new tech crisis (2019)

AI’s white guy problem isn’t going away (2019)

Europe’s silver bullet in global AI battle: Ethics (2019)

A case for critical public interest technologists (2019)

Ethics alone can’t fix big tech (2019)

Government needs an “ethical framework” to tackle emerging technology (2019)

Tech with a social conscience and why you should care (2019)

Trading privacy for security is another tax on the poor (2019)

Congress wants to protect you from biased algorithms, deep fakes and other bad AI (2019)

AI must confront its missed opportunities to achieve social good (2019)

AI systems should be accountable, explainable and unbiased says EU (2019)

One month, 500 thousand face scans: How China is using AI to profile a minority (2019)

How recommendation algorithms run the world (2019)

Facial recognition is the plutonium of AI (2019)

Facial recognition is accurate if you’re a white guy (2018)

Facial recognition software is biased towards white men, researcher finds (2018)

Artificial Intelligence Ideas and Social Work

It is clear that any read of issues of “the future” centers artificial intelligence as a major driver of what is to come. There is a great deal of information out there on this subject. Every profession will likely find themselves impacted by these rapidly changing and expanding technologies – and social work will be no exception. Whether we are directly involved in utilizing these technologies for social good, or addressing social problems that intersect with these tools (and the displacement/or additional social problems they may create), I believe social workers need to learn as much as they can about the power and challenges of these emerging technologies.

AI is and will continue to change the world – social workers must decide for ourselves how best to ethically engage with it as it happens. This curated list is a starting point for social work to learn the lay of the land. This is a HUGE area – so I’m dividing this up into a number of sections.

Fundamentals – what is artificial intelligence?

Here’s the 8 types of artificial intelligence and what you should know about them (2018)

What is artificial intelligence? (2018)

Today’s deep learning “AI” is machine learning not magic (2018)

The non-technical guide to AI (2016)

The jobs AI will create (2017)

Bias in AI

Why AI needs to reflect society (2018)

Racial and gender bias in AI (2017)

Discriminating algorhithms: 5 times AI showed prejudice (2018)

Artificial intelligence has a bias problem (2018)

Using AI for Social Good

Applying AI for social good (2018)

Google using AI for social good guide

AI for Social Good (2017)

Dangers/Concerns of AI

Is AI dangerous? 6 AI risks everyone should know about (2018)

Don’t be afraid of AI (2018)

Limitations of AI

Greedy, brittle, opaque and shallow: The downsides to deep learning (2018)

The real world potential and limitations of AI (2018)

AI and Human Services

AI-augmented human services (2017)

USC Suzanne Dworak-Peck School of Social Work AI Program

All of us in social work need to give the folks at USC Suzanne Dworack-Peck School of Social Work a true bow of respect for the groundbreaking work that they have been doing on this topic INSIDE social work educational and research settings.

Overview of USC AI fellows program

Betting on artificial intelligence to help humanity

USC Center for Artificial Intelligence in Society

Artificial intelligence and social work (book – 2018)