With AI, Sharing Is Caring

Although the public’s distrust of Big Tech is understandable, given past experience, that is not a reason to starve AI models of the information and data they will need to be useful. If serving the public interest is the goal, everyone relying on the technology’s output has a duty to provide high-quality input. NEW YORK—Who wants ignorant AI? Wrong AI? Stupid AI? If privacy and copyright advocates, regulators, and crusaders against the technology succeed in cutting off or poisoning AI’s sources of reliable data and authoritative information, won’t bad AI be the result? My question is an admitted provocation, so we may consider what responsibility media, academics, scientists, companies, governments, and citizens—society as a whole—bear in ensuring that the AI systems that we will all end up using are responsible and credible. To be clear, I do not dismiss the concerns of privacy advocates, copyright holders, or regulators as they address the bottomless maw that is AI, greedily digesting all available data to stitch together its purported intelligence. We must study the privacy implications of hoovering up people’s data—whether online musings and photos or corporate emails and chats—so that both the sources and users of AI data can be protected from unwarranted intrusions. I, too, am appalled at the digital dragnet of Flock cameras and AI’s ability to leverage the data they collect. As an author, I support the value of copyright (though I’ll address my own response to AI ingesting my work in a moment). As with every previous technology, this one calls for sound regulation by the proper authorities based on evidence and experience. I also understand the broad-based resentment against the gargantuan new AI companies and their off-putting, often-obnoxious leaders. It took three decades for the techlash against the internet and then social media to ferment, whereas the backlash against Silicon Valley behemoths, their technologists, and their data centers has reached full, fiery bloom in less than four years (since ChatGPT’s debut in 2022). What troubles me most about AI is less the technology than the technologists and their excessive hubris, capital, and power. The AI boys (which is what most of them are) at the new, so-called frontier labs regularly issue messianic proclamations that they are building a mythical artificial general intelligence with the Promethean power to destroy humankind. Get over yourselves, boys. Continue Reading