Искусственный интеллект рассматривается как инструмент борьбы с конспирологическими теориями во время будущих пандемий

Sign up for the Slatest to get the most insightful analysis, criticism, and advice out there, delivered to your inbox daily. Last month’s Senate grilling of Dr. Anthony Fauci, if viewed by an audience that had not been subjected to years of propaganda, would be staggeringly strange. There were hissing accusations of lies, corruption, and megalomania; comparisons to sadistic Nazi scientists; and statements, as if of undisputed fact, that Fauci killed millions of people. The nation’s most powerful Republicans, without any significant pushback, joined forces to shout down and criminally pursue the face of the COVID pandemic’s health authority based on wild, speculative, and facile conspiracy theories. It all prompted a kind of deeply existential question: If this is how we now treat a onetime pandemic hero, what will happen the next time we face a similar public health crisis? Public health officials say with the rise of conspiracy thinking at the highest echelons of our national government, the next pandemic response could be so much worse. But they did have one potentially innovative and surprising, if disconcerting, solution to our national public health information crisis: A.I. has the potential to help stave off a conspiratorial catastrophe during the next pandemic. Given everything that’s happened with our public health leadership since Donald Trump returned to power, most public health professionals Slate spoke with took a dim view of our possible responses to a future crisis. After a summer of foodborne illness crises, it’s clear our public health officials are struggling to handle basic public health. Indeed, they seem to be taking us backward: Just last week, it was reported that school vaccination rates have hit a new record low. Still, how worried should we be about another pandemic? Historically, pandemics have happened multiple times a century, and many scholars believe that they’ll become more frequent due to climate change and deforestation. In the six years since COVID arrived in the U.S., the public health community has drawn no significant new conclusions about how to beat back conspiracy theories during a pandemic. The situation appears bad enough that when Slate asked experts how to conspiracy-proof a pandemic, responses often offered such bleakly unhelpful tidbits as “We need to study the effectiveness of our communication” and “This is going to take careful thought about strategic communications.” As Tara Sell, a senior scholar at the Johns Hopkins Center for Health Security, told Slate, “We’re in a bad spot right now.” But several experts, prodded for any note of optimism they could find, offered hope from artificial intelligence. Admittedly, looking to A.I. for rescue is usually an eye-roll-worthy impulse. And coming from the field of disinformation in particular, A.I. is a Pandora’s box of yet-to-be-imagined evils. Generative A.I. certainly has the potential to worsen pandemic conspiracy theories by equipping them with highly convincing fabricated evidence. But in this case, the desperation of the situation might just call for public health advocates leaning into the dangerous new technology. According to Tim Mackey, a professor of global health at the University of California, San Diego, A.I. responses appear to be getting better at summarizing research, weighing credibility, and discerning misinformation. This applies only to the A.I.-generated search results summary and not to chatbots, which are inclined to give users the answers they want. But for a single search, the A.I. performs “decently,” Mackey said. “And that’s better than social media.” It’s also better than straight Google searches, in which users “doing their own research” tend to click the links that will support their views and interpret the information they find how they want to, potentially reinforcing conspiracy theories. The hope is that A.I. will help by encouraging laziness—the less time we spend pursuing answers, the better. Of course, Mackey said, there’s no guarantee that the A.I. won’t go off the rails. If A.I. companies got into certain commercial models, they could start promoting sponsored results, such as pseudoscientific cures. Or they could bend to pressure from conservative activists or tech oligarchs who see the weeding out of conspiracy theories as the suppression of political speech. So the possibility that A.I. might help is admittedly a flimsy source of comfort. But it’s something. There’s a second theory that might point to A.I. as our savior, though. Multiple experts Slate spoke with agreed that deplatforming the conspiracy theorists didn’t help; purging bad actors from platforms only made things worse, they said, as the people feeling censored hardened in their views and audiences followed those accounts to even more insular and unregulated platforms. Some experts mentioned other technological solutions to help with public health: namely, the tweaking of algorithms to make it so the most inflammatory accounts aren’t always rewarded. But one scholar had a more offense-oriented idea. Neil Johnson, the head of the Dynamic Online Networks Lab at George Washington University, sees the problem of conspiracy theories in health emergencies as the natural result of the huge scale of online communities. Before COVID even had a name, he said, online parenting communities on Facebook were reaching out to other communities to get information about the new flu-like virus in China. They found ready information in anti-vax groups, which claimed to be able to answer questions that the public health experts could not; conspiracy theories will always move faster and more decisively than legitimate scientific inquiry. And because of the points of connection between different Facebook groups with shared interests, the anti-vax ideas spread like wildfire, reaching millions of users and, by extension, their loved ones. The scale of the problem is so vast, Johnson said, that it requires an equally ambitious strategy. And his idea is this: Have tech companies use A.I. to spur interactions among conspiracy theorists, pro-science activists, and the middle-ground groups, such as the parenting communities. Studies show that when you put all three together, he noted, the extremes tend to soften. “And over time, it will soften the entire universe,” he said. It’s a grim idea, that our hope is in a tool that discourages in-depth reading and a long-shot hypothetical strategy—both of which rely on technology known for sowing confusion. But consider what the public health communicators are up against. The public is primed to believe conspiracy theories because they don’t trust the medical system, in part because we operate in a fragmented and highly confusing for-profit system. Many people do not have primary care physicians, and often those who do cannot easily call up their doctors to ask questions. We don’t know who to go to for trusted information. These are systemic problems without easy solutions. Pandemics worsen the suspicion. They are fast-paced, overwhelming, and chaotic, and public health officials inevitably make mistakes. Responses involve multiple agencies and levels of government, and because of our polarized environment, any information coming from a presidential administration will be met with automatic suspicion by the president’s opposing party. And, at a more fundamental level, pandemics breed conspiracy theories because someone inevitably loses. Once the disease gets out of hand, someone has to take a lower priority: elderly people vulnerable to illness versus children missing out on school, for example, or business owners facing closures versus hospitals worried about limited beds. Those who study conspiracy theories hope that the next federal pandemic response will do a better job of communicating uncertainty. They also hope it will not be embodied by a single man, a central messenger who, like Fauci, can be easily discredited in certain communities. They hope the arguments within the community will be less pedantic. (Who cares if we say aerosolized or airborne?) That people will acknowledge the pain and sacrifices that come with public health measures. That community leaders and representatives of the business and faith communities will feel included in the decisions. (Unsaid in this is that public health officials have to be comfortable relinquishing control and allowing more people to get sick or even die in the short term in exchange for long-term compliance.) That we will not try to drive a one-size-fits-all response across the states. But none of those improvements will stop conspiracy theories. Nor will media literacy programs, stronger education about viruses, centralized hubs for medical information, “prebunking” campaigns, or community engagement programs—all good ideas from experts that still cannot meet the sheer force of our paranoia. The conspiracy theories, they all agree, will still happen. The distrust is too bad, the hucksters too loud. If it were possible to conspiracy-proof a pandemic, we would get there only with investment and strong leadership. Instead, the federal government has eliminated positions that dealt with “infodemics” and slashed funding for research into disinformation. Academics are afraid to probe questions the administration doesn’t like, worried about grant money. Our top health official is a COVID conspiracy theorist who has staffed the agencies with other conspiracy theorists. The entire federal government is scaling back on providing services to the public. This all leaves us with the pathetic situation of working off a pandemic strategy developed before the current social media age, hoping the American public will experience a radical change in attitude, and turning for comfort to the whims of artificial intelligence. “We’re just trying stuff and seeing if it sticks,” said Jennifer Nuzzo, the director of the Pandemic Center at Brown University, when discussing the conspiracy theories. “But we’ve barely scraped the surface in terms of understanding.”