Uncanny Valley: Navigating the Tangled Web of AI, Privacy, and Digital Lives
Welcome to Uncanny Valley, your weekly deep dive into the stories shaping our increasingly digital world. Hosted by WIRED’s Director of Business and Industry, Zoë Schiffer, and Executive Editor Brian Barrett, this episode unpacks a diverse range of critical developments, from the intimate confessions whispered to chatbots to the sweeping data collection practices of government agencies. We’re navigating the uncharted territories where technology intersects with our personal lives, our security, and the very fabric of society.
China’s Digital Divide: Apple and the Shifting Sands of Online Expression
Our journey begins in China, where the digital landscape continues to be shaped by government directives. WIRED colleagues Zeyi Yang and Louise Matsakis have reported on Apple’s recent decision to remove two prominent gay dating apps, Blued and Finca, from its Chinese App Store. This action, prompted by an order from China’s primary internet regulator, signifies a significant blow to the LGBTQ+ community in a country where options for connection are already limited and same-sex marriage remains unrecognized. While existing downloads may still function, the removal curtails access for new users and underscores the delicate balance tech giants must strike to operate within the Chinese market.
Brian Barrett highlights the rarity of Apple confirming such an order, noting their typical reticence regarding their operations in China. "It’s all a pretty live wire for them," he observes, referencing the complexities of data handling and supply chains in the region. The situation also brings to mind Apple’s recent decision to remove ICE (Immigration and Customs Enforcement) detection apps from its stores, a move that drew parallels to conforming to prevailing political climates, even when it seems to contradict foundational principles.
Zoë Schiffer points out the broader implications: "If you are going to be in the Chinese market, you need to share data with the Chinese Communist Party. You need to comply with their core orders, and those can change on a dime it seems." This reality forces companies to confront the prioritization of business interests over stated values, a lesson that has become increasingly evident in recent years.
Data Centers: Powering the Future, Powering the Planet
The exponential growth of artificial intelligence and cloud computing has placed data centers at the forefront of technological infrastructure. But as WIRED’s Molly Taft reported, the critical question isn’t just if we need more data centers in the United States, but where they should be built to mitigate environmental impact. A new study published in Nature Communications Journal offers a compelling framework for sustainable data center placement.
The ideal locations, according to the research, must meet two key criteria: the local electricity grid should be powered by or making significant strides towards renewable energy, and the region should not be experiencing severe water scarcity. Based on these factors, states like Texas, Montana, Nebraska, and South Dakota emerge as prime candidates.
Brian Barrett finds this analysis particularly insightful, noting the potential for states like Nebraska to leverage wind energy to power data centers. "If you want to build out this huge AI infrastructure, you’re going to have to embrace some of the renewable energy that a lot of this administration has obviously turned its back on quite a bit, but you need that," he emphasizes. This underscores a crucial dependency: the ambition of AI development is inextricably linked to the adoption of sustainable energy practices.
Zoë Schiffer shares a firsthand observation from a visit to an OpenAI data center in Abilene, Texas. Despite the push for renewables, the reality on the ground involves the use of massive gas turbines, a testament to the insatiable energy demands of these facilities. "We need everything," she quotes the CEO of the company managing the buildout, highlighting the imperative for uninterrupted operations.
The conversation also touches upon the growing public resistance to data center development. "It’s going to be politically untenable to be supportive of having a data center in your local community, because no one wants to live near these things," Zoë predicts, citing concerns over noise, energy bill spikes, and pollution – classic NIMBYism dynamics.
Brian adds another layer of concern: the limited job creation offered by these massive infrastructure projects, often coupled with significant tax breaks. This imbalance between economic benefits and local impact is fueling growing pushback.
The Scourge of Scam Texts: Google Takes on Lighthouse
In an era of constant digital connectivity, scam text messages have become an inescapable nuisance. "It is incessant," Brian Barrett exclaims, echoing the sentiment of millions who receive a barrage of unsolicited and often fraudulent communications daily.
WIRED’s Matt Burgess has investigated Lighthouse, a sophisticated Chinese fraud network responsible for millions of scam texts annually. Impersonating entities like USPS or toll road collectors, Lighthouse has reportedly amassed over a billion dollars through its schemes. In a significant move, Google has filed a lawsuit in the United States against 25 unnamed individuals allegedly operating within this network.
The lawsuit targets the operation’s core: Lighthouse sells a "scamming as a service" software, which even provides subscription models for less tech-savvy scammers. This professionalization of fraud, from weekly to permanent subscriptions, is a chilling development.
Brian notes the recurring pattern of "service" models in illicit online activities, from ransomware to now, scamming. He also acknowledges the potential futility of such lawsuits, particularly when pursuing perpetrators in different jurisdictions. However, he highlights the downstream effects: "At the very least, if you can prove that these people are doing this in court then you can maybe get some traction with the platforms they’re working with… to cut off avenues for them to operate."
Zoë finds it telling that while individuals aren’t named, their Telegram handles are included in the filing, hinting at the digital footprints these scammers leave behind. Despite their billion-dollar success, Zoë humorously questions the persuasive power of the phishing templates, suggesting perhaps the scammers’ revenue speaks louder than her skepticism.
AI Romances and the New Frontier of Divorce
As artificial intelligence becomes more integrated into our lives, the nature of human relationships is also being redefined. A groundbreaking story by WIRED’s Jason Parham explores a new and complex legal territory: AI affairs as grounds for divorce.
"So much to unpack here," Brian exclaims, acknowledging the increasing prevalence of individuals forming deep emotional bonds with AI companions. The legal implications are profound, particularly in states requiring fault for divorce. Is an AI affair considered infidelity? Is it an addiction? These questions pose significant challenges for individuals and legal professionals alike.
Zoë empathizes with the emotional toll, suggesting that an intense texting relationship with a chatbot could indeed feel like an affair, albeit a non-traditional one. The financial implications are also significant, with potential impacts on asset division in cases of financial waste through hidden payments or subscription fees for AI companions.
Custody battles could also be affected, as such relationships might raise questions about a party’s judgment. Brian finds it difficult to envision a scenario where a party could successfully argue for custody while simultaneously being in a relationship with a large language model.
OpenAI and the Battle for AI Privilege
This discussion of AI relationships also brings to mind the ongoing lawsuit between The New York Times and OpenAI. OpenAI’s recent filing argues against being compelled to reveal thousands of user conversations with ChatGPT, proposing a concept of "AI privilege." The company contends that the deeply personal, emotional, and even professional conversations users have with AI should be protected, akin to attorney-client privilege.
Brian expresses skepticism, viewing this as part of OpenAI’s ongoing narrative of exceptionalism. He questions the validity of such a claim, especially when it potentially hinders legal proceedings, such as divorce cases involving AI affairs.
Zoë sees merit on both sides, acknowledging the deeply personal nature of AI conversations compared to typical search queries. However, she questions whether this distinction warrants a unique legal protection. The debate raises fundamental questions about data ownership, user privacy, and the evolving legal landscape surrounding AI.
Furthermore, OpenAI CFO Sarah Friar recently admitted to a cooling growth in ChatGPT usage, partly attributed to content restrictions. The company’s stated intention to roll back some restrictions, including allowing erotica, signals a strategic move to re-engage users and drive growth by catering to more intimate and romantic interactions. This shift is met with concern from Brian, who worries about the incentives driving OpenAI’s development trajectory.
The DHS Data Breach: A Violation of Trust in Chicago
Our main story this week centers on a disturbing privacy breach involving the Department of Homeland Security (DHS). WIRED’s Dell Cameron has reported that the DHS illegally collected data on hundreds of Chicago residents, supposedly for testing purposes to see if police files could feed into a federal watch list.
The data, originating from a private exchange between DHS analysts and Chicago police, was intended to explore how local intelligence could enhance federal watch lists. The core objective was to determine if street-level data could identify undocumented gang members and flag them at border crossings.
However, the data itself was deeply flawed. Internal memos revealed that the dataset was initially requested in the summer of 2021, despite Chicago’s gang data being notorious for its inconsistencies and lack of robust maintenance. The report found that individuals on the list weren’t even confirmed gang affiliates, with police officers often including their personal opinions, leading to entries like "scumbag" or "turd" as occupations. A stark and troubling statistic emerged: 95 percent of the individuals on this list were Black or Latino.
Zoë highlights the particularly severe implications for immigrants in Chicago. Due to sanctuary city rules that limit data sharing with immigration officers, an exception for "known gang members" created a loophole. Over a decade, immigration officers accessed this database more than 32,000 times, often as a workaround.
Brian details the lack of oversight, with the data moving through layers of review without a clear owner or legal safeguards. The project was ultimately terminated in November 2023, with no clear tangible results. Yet, the potential for harm is immense, as sources indicated that such breaches could impact millions, potentially justifying federal targeting of families, churches, and food banks that support immigrants—even those without any gang affiliation.
"There are so many layers of inappropriateness to the use of this data and the gathering of this data," Brian states, emphasizing the sloppiness of the process and its severe real-world consequences.
The current administration’s approach to immigration, characterized by a willingness to use any pretext to deport individuals, exacerbates these concerns. The DHS’s data gathering ambitions, coupled with an executive order aimed at eliminating data silos, raises alarms about the consolidation of sensitive information and its potential for misuse. As Brian observes, "there is such a push to gather as much data as you can about as much people as you can… and there’s really no regulatory framework that would prevent any of this."
Zoë concludes with a sober reflection on her brief experience in the US government, noting that while nefarious intent isn’t always present, bureaucracy, laziness, and sloppiness can lead to equally devastating outcomes. The DHS data breach serves as a stark reminder of the constant vigilance required to protect our privacy in an increasingly data-driven world.
Stay tuned for next week’s Uncanny Valley, where we’ll delve into the world of Palantir CEO Alex Karp.
This episode was produced by Adriana Tapia and Mark Leyda. Mixed by Amar Lal at Macro Sound. Executive Producer: Kate Osborn. WIRED’s Global Editorial Director: Katie Drummond.