Faith, Technology, and “the Powers”

An aura of inevitability tends to surround artificial intelligence. Barely two months after OpenAI launched ChatGPT in 2022, powered by OpenAI’s large language model GPT-3, the chatbot gained 100 million active users and set off a race among the large technology companies to release their own models. 

The probabilistic nature of AI “decision-making” explains why AI is not concerned about truth per se but rather about reducing margins of error.

The U.S. Department of Defense is working on the possibility of using AI-enhanced autonomous “swarms of swarms,” massive throngs of drones as weapons of warfare to address China’s emerging military power. China is investing heavily in AI in part because it believes AI is vital to countering what it believes to be the superior military force of the U.S.

AI is now at an inflection point and stands to affect every domain of human life. As AI grows in power, it is worth considering the extent to which it is indeed inexorable and to seek what wisdom our faith offers as we respond to it. 

Demystifying a Juggernaut

AI is a set of computer programming techniques used to perform relatively sophisticated tasks. Yet they are only programs, often trained with a vast amount of data, and as such their “experience” of the world is mediated through numbers. The aim is to make their best guesses at solving a given task, with greater accuracy and specificity each time.

This should dispel some of AI’s mystique. The probabilistic nature of AI “decision-making” explains why artificial intelligence is not concerned about truth per se but rather about reducing margins of error. Legal scholar Ryan Calo points out that AI remains dependent on human beings: besides computer programmers, a company often contracts with data suppliers who hire hundreds of people, often in poorer countries, to label datasets, sometimes under arduous conditions. Large language models are trained not only with data scraped from the internet, but also necessarily by people who interact with the programs and who correct their output.[1]

Contradictory Impressions

Still, it would be naïve to ignore AI’s expanding uses or the strong incentives to improve and deploy it. AI applications could be highly beneficial to humankind. At the same time, they will exacerbate inequalities, primarily because of uneven access to technology, capital, and education. They are likely also to amplify harms and create new ones. But the prospect for good and the imperatives of the market and of international relations seem to assure AI’s continued development.   

These trends reflect tensions that run through society’s competing conceptions of technology in general. Technology is viewed as unstoppable. Or as neutral. Or as embedded in politics. Proponents of technological determinism argue that tech breakthroughs are driven by the desire for knowledge and the capacities that technology affords. Moreover, technology is hard to control: the Collingridge dilemma posits that when a technology is in its early stages, it is too soon to regulate because it is impossible to predict what effects it will have, but by the time those effects are known, the technology is too widespread to harness. 

Others insist, however, that technology is neutral: outcomes depend on the user. Thus, to the extent that any controls are placed on it, they should regulate the user’s goals and behavior, not the technology itself. Finally, some contend technology cannot be understood apart from its political aspects, not only because it is deployed for social control, but also because it depends on political decisions and human institutions to exist. 

All three accounts have a grain of truth to them, and all three make their way into the rhetoric that accompanies debates about AI. What then are people of faith to make of these dynamics? What insights might the scriptures provide as we work out the implications of the gospel in relation to what seems to be an AI juggernaut?  

“This Present Darkness”

The author of Ephesians writes that “our struggle is … against the cosmic powers of this present darkness, against the spiritual forces of evil in the heavenly places.”[2] Biblical commentators such as Beverly Roberts Gaventa argue that the New Testament writers assumed that the principalities and powers are “real.”[3] The scriptures do not “provide … a systematic account of their origin or status,” but they are not “merely code language for the empire or its agents. Nor are they to be internalized in social or psychological terms.” “To be sure,” Gaventa says, “we may well see or suspect the activity of the principalities and powers in human empires or within ourselves, but they cannot be reduced to such.”[4] Theologian Paul Hinlicky similarly cautions against mischaracterizing the powers. As the New Testament describes them, they remain a mystery—an inexplicable force of malevolence—and a matter to be discerned with care by communities of faith.[5]

We might pursue this approach as we assess AI and the actors and circumstances that are shaping it. Scripture recognizes that large, complex systems often figure into the harms people suffer. According to Paul, when the community of faith discerns the powers at work, it should not submit to them.  Instead, the church should pursue the life made possible by the gospel, confident that God transcends all powers and will allow none of them to separate us from God’s love.[6] At the same time, the texts warn us against equating too easily the powers with any specific structure, institution, personality, or as we could say now, a nascent “intelligence,” if for no other reason than so our actions will not be misguided and, crucially, so that we will not be held in thrall to them. 

Christ and Culture and Niebuhr

The New Testament writers understood that forces beyond any one individual or community have a decisive impact on their lives. That certainly resonates with our times. But they were not cowed by those forces because of what God had done in Jesus Christ. This frees us to demystify them so that we can promote them when they contribute to our communal life and oppose them when they do not.

Our response to the powers will also be informed by the way believers feel called to be in the world. Although there are no ideal types, H. Richard Niebuhr showed long ago how faith communities, based on their diverse understandings of Christ’s relationship to culture, have associated with the world in different ways, from seeking almost complete separation to making no distinction between themselves and the secular.[7] Views of technology have fallen roughly along this spectrum, from decrying technology as hubris to celebrating Jesus as the ultimate artisan, the tektōn.[8]  

In any case, to use language from Jeremiah, even a community that perceives itself as a people in exile is called to seek the peace and prosperity of the places to which they, with their gifts and skills, have been carried.[9] Such a call to promote and honor human flourishing, as well as maintain freedom from thrall to the powers, invites us to work with others who are trying to steer AI in directions that benefit humankind. 

Contentious Attempts at Control

As part of these efforts, a “law” of artificial intelligence comprising a range of practices, norms, and formal legislation is emerging from the interplay, sometimes cooperative and sometimes contentious, of communities, businesses, governments, and other stakeholders—interactions at all levels of governance, locally and globally. 

In March 2024, for example, the European Union formally adopted the Artificial Intelligence Act, which marks a major development: the Act’s primary purpose is to ensure that AI systems are used consistently with human rights. Parts of the legislation apply directly to large language models as well as to other forms of generative AI. In the area of peace and security, countries are negotiating a protocol to regulate lethal autonomous weapons. Efforts to ban them are opposed by several of the major powers, in the name of supporting their national security goals. That opposition, however, has not kept other stakeholders and some 30 nations from insisting on such a ban. Further, all participants in the process acknowledge that the use of AI in warfare is constrained by humanitarian law.

When to Support, When to Oppose

The scriptures meanwhile caution us to be clear-eyed about the prospects of success of such proposals, given the powerful actors and forces at play and, recalling Collingridge’s dilemma, the limits of the tools we have for governance. Because of the interests of some countries, including the United States, prohibition of lethal autonomous weapons—now in the tenth  year of discussions—looks unlikely for now, and we are well aware of the inability of humanitarian law at times to prevent widespread destruction. As for the new EU law, only time will tell just how effective it will be in regulating AI systems.

Despite these machinations and the seeming frailty of our efforts to control them, we are reminded that God’s strength is made perfect in weakness.[10] What practical steps therefore might a congregation, denomination, or individual take if they want to become more involved in these matters?[11] The first is to learn more about AI and its potential uses, identify an area of particular interest or urgency, and consider supporting organizations dedicated to reform and advocacy. Trends of AI in the classroom, in healthcare, and in eldercare are just a few of many such areas.

All of this suggests that for communities of faith, our decisions about when we are to oppose AI or when we are to support it should be guided by our call to cultivate human flourishing and our trust that God will honor our efforts to further it.

Mark Chinen ’84 M.Div., a professor at the Seattle University School of Law, is a member of the organization AI and Faith, which advocates that the uses of technology be exclusively life-affirming and promote human dignity and a just society. He is the author of The International Governance of Artificial Intelligence (2023) and Law and Autonomous Machines: The Co-evolution of Legal Responsibility and Technology (2019), both issued by Edward Elgar Publishing.