Placing bars on the windows of a house can make it more difficult for others to break into and perpetrate intentional harm. Placing bars on windows does not fill a home with joy and possibilities. Similarly, we know that guarding against harm in AI is not equivalent to maximizing the good it can produce. Only people guided by principles, not rules, can maximize good. This distinction is sometimes lost. The White House’s recently released guidelines, to which AI leaders voluntarily agreed, unfortunately do not guide AI towards maximum good; however, faith teachings do guide us towards such maximum good. Much of this guidance is now more accessible to (secular) society and business through the field of Organizational Development (OD). In the Western world, such principles are paralleled, or perhaps originated, in Christian scriptures – such as 1 Corinthians 12:20 and Acts 4:32 – that implore adherents to orient towards collective well-being.
History suggests that it would be errant to limit our collective imagination only to concerns that we currently anticipate – AI architects are currently contemplating how to address increased energy consumption and the amplification of human biases. We will produce maximum good when guided by organizational development principles.
AI and other emerging tools such as blockchain will again reshape societies, just as social media has during the past two decades. Undoubtedly, AI will produce many new capabilities, some currently seen and many more unseen. It will solve currently intractable problems such as protein folding to improve medicine, enhancing automobile safety, and personalizing learning – while likely creating new problems to be solved in the future. Given the certain scale of impacts, AI architects must employ organizational development principles to increase AI’s contribution to the collective good.
Known Harms
Earlier information sharing innovations such as the printing press, telephone, the internet, social media, and even the automobile each reshaped society. The automobile provided the ability to travel further faster. In the unanticipated realm, it also facilitated the dispersion of families, reduced the walkability of cities, dissected communities, and accelerated carbonization of the earth’s atmosphere.
Similarly, social media, one of the big technologies of the past two decades has produced many benefits along with significant unintended consequences. They include reduced self-esteem among teenage girls, that became more visible as the technology scaled. Research has long shown that social media use increases rates of depression, anxiety, low self-esteem and eating disorders among communities of girls and young women. The neuroscience here is mature. This knowledge did not stop the social media companies from continuing to extract loads of money in the short term. The long-term damage is a form of disconnectedness that has social consequences beyond ad revenue for the tech industry.
Many have already illuminated the fact that algorithms amplify assumptions and biases in ways that magnify existing human harms. In 2020, Dr. Abeba Brhane’s discovery of substantial bias and harm perpetrated by a then “highly regarded” MIT data set provides just one such example.
For over a decade, the data set containing over 80 million images was fed into many machine learning models to train artificial intelligence systems. Researcher Abiba Birhane found that the MIT Tiny Images data set was rife with demeaning gender and racial hostilities, including private photos taken of hundreds of women absent the subjects’ knowledge. With impunity, data scientists cited the substantially impure data in hundreds of academic papers. The hubris and homogeneity of the profession shielded the data’s fidelity from inspection. To the contrary, such foundational data only served to accelerate the demise of the America’s social fabric.
Beyond bias, even more powerfully, AI’s ability to invent new narratives and its power to execute decisions without human involvement amplify its potential impact.
Starting with the End in Mind
Maximum collective good requires solutions which treat economies as a means, rather than an end. A strong economy builds social connectedness – it does not erode it. In turn, the foundation of social connectedness requires that we view all – including those we have long feared and othered – as deserving human dignity.
Though not an AI or even an information technology enterprise, Costco serves as an alternative example to Meta’s extractive dehumanizing principles, recurring harm, and under-fulfilled social promises. Costco has produced lasting strength for itself and the communities it interfaces with, stemming costly recurring social and economic harms that many industries ignore.
My field of organizational development is most often employed to support alignment during times of significant business restructures. Whether restructurings are driven more by external shifts or business model innovations, OD practitioners aim to align strategy, operations, and culture. The foundational principles of collective well-being, systems integration, and partnership health could guide AI towards better workplace cultures, products, and supply chain partnerships. Beyond guarding against specific harms such as those imposed by biases, these guiding principles will advance the common good. They serve as positive catalysts, encouraging power sharing and reducing opportunities for systems to propagate future unintended harm.
-
-
- When collective well-being is prioritized in a society, everyone lives with greater dignity. More individuals retain a sense of belonging and actively engage in their communities. Such dignity requires that conditions sustain individual agency, strengthen communities, and prioritize economic equity. To ensure such outcomes, AI systems must include associated parameters and incentives in their reinforcement learning as systems are being trained. OD professionals know that while designing new business models that will invariably eliminate some jobs, the new model is enhanced when current frontline employees in the retiring model play a significant role designing the new. Such engagements, absent protectionism, invariably produce healthier products for all stakeholders. This is part of what entertainment industry writers are lobbying for.
-
Had early civic transportation designers considered not just the expeditious benefits of the automobile, but some of what many loved about their communities, we might currently be enjoying some of the unique benefits of the automobile while spending fewer public resources restoring open spaces, encouraging children to engage in more physical movement, and mitigating automobile pedestrian fatalities among the elderly. The challenge for such responsible enterprises will be how to build in well-being in the near-term as some competitors prioritize only the more tangible new capabilities.
- Whether a membership warehouse, airplane manufacturer, or healthcare, all supply chains demand better systems integration. The needs of all stakeholders must be balanced. In healthcare, medical research confirms connections between social trauma and illness. An often-cited the 1998 study by Kaiser Permanente and the US Centers for Disease Control and Prevention documents the fact that adverse childhood experiences such as abuse and neglect play a major role in physical disease and mental illness and can be transmitted across generations. As a result, effective treatments for substance abuse and domestic violence, among other problems, increasingly consider larger parts of the system. Among Indigenous communities, cancer treatment encourages traditional diets and traditional ways of gathering and preparing food, family, and community, as well as spiritual and social dimensions.
- Like the branched human vascular and nervous systems, artificial neural networks are fractal, meaning strengthened when designed to be reinforcing at each level. Currently, organizations such as Costco that provide employee ownership at each level are proving to possess durable growth, brand loyalty, high levels of employee engagement, and supply chain resilience. AI architecture will produce greater social well-being when upstream and downstream partnership health are considered from the start. In addition to business flow, such partnerships can include the perspectives of communities that have been historically feared, othered, and those that prioritize ecological balance.
Beyond harm reduction, American business knows the long-term value created through alignment such as vertical integration, globalized networks, and President Biden’s current industrial policy.
AI is not the only nascent technology we must cultivate well by applying OD principles of complex systems. AI will mature together with blockchain, the metaverse and possibly other powerful computational technologies. American’s past dynamic economic eras including industrialization and financialization each amplified foundational national “design flaws” engineered from the beginning to exclude women and non-European men. Such past ascendant industries also generated substantial externalities, such as income stratification and environmental degradation that continue to burden taxpayers long after individual corporations and industries fade into obsolescence. These foundational flaws excluded concern for the dignity of many who held valuable talents and perspectives.
To maximize America’s collective well-being and reduce future unsustainable social repair costs, AI and other ascendant technology industries must now engineer outcomes that prioritize the full potential of all who call the nation home.
