In his 1955 short story Franchise, Isaac Asimov imagined how American democracy might be radically transformed by the digital age. In the story, set in 2008, Americans’ political will is exercised not by individual citizens who stand in line to vote, but by a massive supercomputer—the Multivac—that processes an ocean of public data with inscrutable algorithms to reliably predict the outcome of this messy, partisan, costly, and all-too-corruptible process.

The story works on many levels, but above all it evokes the technocrat’s dream (or dystopian vision, depending on one’s perspective) of using new technology to smooth out the wrinkles in our aging analog democracy.

Today, those wrinkles are looking more and more like cracks, while new technologies—from social media and predictive search to digital surveillance—seem to be doing more to destabilize our fragile democratic institutions than to reinforce them. What can we do? I suggest we look at why our newest digital tools—until recently, celebrated as bearers of a new age of democratic wisdom and civic health—have largely failed to deliver on those promises. The causes are complex, and any brief account of them doomed to inadequacy. But chief among them is the technocrat’s error of building ever more powerful systems and platforms for democratic life, while entirely neglecting the need to cultivate robust civic habits, norms and virtues among the peoples who will use those systems.

Technology’s threat to democracy is not, at its root, that of poorly designed systems (though certainly design improvements can be made). The real threat is when technical progress is relied upon as a substitute for moral progress in cultivating the civic virtues, norms, and values that sustain functional democracies.

At this year’s RSA information security conference, Eric Schmidt, the chairman of Alphabet, attributed the recalcitrant challenge of Internet security to a technocrat’s blind spot: in designing internet platforms and protocols that would eventually be vulnerable to hackers, he said, “it didn’t occur to us that there were criminals.” If we ask why it has been so difficult to civilize Twitter, or to weed out falsehoods and hoaxes on Facebook, one might likewise reply that those systems were not designed with sufficient concern for the existence of pathological liars, demagogues, dogmatists, racists, sexists and other species of civic vice. More to the point, the designs of such platforms have assumed civic virtues as inputs, rather than helping to cultivate them—virtues like integrity, courage, empathy, perspective, benevolence, and respect for truth necessary to fuel any democratic technology, analog or digital.

Regrettably, American public education has itself largely abandoned its civic mission, and reform from that quarter is unlikely to come any time soon under present leadership. In the meantime, companies like Twitter and Facebook—whose futures depend upon democracy’s health—should prioritize and fund intensive study of the civic norms, skills and virtues of their users, and how these can be better expressed, refined, and fostered by specific platform designs and features. The threads of technical and civic progress can and must be rewoven.

This article is part of a collaboration with the Markkula Center for Applied Ethics at Santa Clara University.