With AI, Americans Aren't Being Asked To Trust but To Believe

With grassroots resistance to AI data centers growing, Trump last week urged localities to embrace them. Because one reason for public opposition is higher utility costs for households near centers, Trump gestured toward a Ratepayer Protection Pledge his administration announced last winter. He badgered local leaders: “You have to convince your community how great these things [are],…You can’t fight it. You have to go with it, and you’re going to have all that protection.”

What exactly is “all that protection” provided by the Ratepayer Protection Pledge? The Pledge isn’t regulation. It’s a document with various corporate and government signatories that just declares data centers will pay for their own energy use. It holds no one accountable and makes nothing enforceable. It is less than a promise. It is a prayer.

Trump’s use of the AI inevitability argument (The gods are omnipotent!) sloppily wrapped in an appeal to trust (The gods are benevolent!) reminded me of a recent paper by an interdisciplinary team of Australian and Dutch scholars led by Deakin University’s Bianca Baggiarini. Their compelling and urgent work, “A ‘journey to trust?’ Machine learning, faith, and the discourse of ‘trusted’ autonomous systems,” focuses on military adoption of AI.

The authors argue that the rhetoric of trust around AI encourages its adoption as it turns the discussion from whether the technology and its purveyors can be trusted to whether and how trust in them can be built:

Many ethicists and legal scholars agree that to ensure lethal autonomous weapons systems (LAWS) – i.e. systems which often rely on complex algorithms and large datasets that are not fully transparent or understandable to its users – are acceptable, they must, above all else, be trusted. Given the current fixation on trust (Troath 2024, 2) significant research is being dedicated to enabling trust…We claim that trusted autonomy discourse helps accelerate technology design and deployment while eclipsing questions related to power and politics (Troath 2024). Here, trust is regarded as a ‘value to hold, and a problem to solve’ (Troath quoting Richardson 2022).”

When talk turns to how stakeholders can come to trust AI, skepticism, rather than the technology, becomes the problem. The goal and result isn’t greater public trust, which depends on honesty, respect, and transparency. The authors explain: “‘trusted autonomy’ discourse serves as a smokescreen for faith, legitimizing the values and perspectives of military elites and technology developers.”

Though this paper is focused on the military, legitimizing research on trust in AI and other automated systems is being conducted across fields and sectors. When AI proponents ask, "How can we gain trust?", it can sound as though they intend to investigate potential downsides and side effects, work on improving product safety, or accept regulations or restrictions. But what purveyors prefer is to build a flock of the faithful.

Edtech vendors, for example, are changing their messaging to school districts facing parent pushback on screen use; rather than research their products’ effectiveness, as Emily Cherkin notes, they’ve studied and coopted the language of skeptics who question their purposes. Harvard Business Review offers research-based classes on how managers can increase workers’ faith in AI tools. The praiseful course description serves to bolsters bosses’ belief in AI’s ability to accelerate productivity and move their firm forward — while helping them make converts of their employees.

Baggiarini and her co-authors note how conversations around “trust” reinforce the idea of AI’s inevitability: “Trust discourse urges ‘us’ – citizens, soldiers, and states – to believe in AI’s promised future potential and to accept its inevitability in war, even as many witness their well-documented failures, risks, and harms.”

The tech industry prefers to encourage faith rather than establish trust. The nature of AI systems — complex, opaque, rapidly-evolving — means a leap of faith is required for implementation, especially when it comes to lethal military applications. Encouraging faith benefits AI vendors, though, whether they are selling to schools, businesses, homes, or government agencies since building trust is so “time-consuming” and “messy.” Fortunately for AI vendors, many people already worship at the altar of technology, as anthropologist Catherine Lutz and I have written in the context of car culture:

If there is one true religion in the US, it leads us to worship at the altar of technology…we accept the doctrine of this shared faith: that technology provides the main path to improving our lives and that if it occasionally fails, even catastrophically, it will just take another technology to make it all better.

Tech industry leaders have long benefited and enormously from this faith, and as the elite of America’s elite, they understand (as does their beneficiary Trump) that faith provides a pathway in this era of distrust. It’s easier and faster to chant, “Behold this almighty power.”