23AR28-34

( - previous issue - / - next issue - )

pdf = www.tinyurl.com/2s3dctb5 (corrected issue number)

AR 28:34 - US church membership falls behind population growth

In this issue:

AMERICAN CHRISTIANITY - Protestant denominational growth worse for evangelicals than expected

ARTIFICIAL INTELLIGENCE - a cat-and-mouse scramble continues on into the 21st century

Apologia Report 28:34 (1,631)
September 27, 2023

AMERICAN CHRISTIANITY

"Just How Bad Is Denominational Decline?: A deep dive into the trajectory of nine denominations" by Ryan Burge (Graphs about Religion, Jun 12 '23) -- First of the Protestant denominations considered are "six in the mainline tradition: American Baptist Church, Evangelical Lutheran Church of America (ELCA), Presbyterian Church USA (PCUSA), the Episcopal Church, the United Church of Christ (UCC), and the United Methodist Church (UMC). Those six represent the vast majority of the mainline.

   Three in the evangelical tradition are also included: the Assemblies of God, the Southern Baptist Convention (SBC), and the Presbyterian Church in America (PCA). "The SBC is easily the largest denomination in the evangelical family and the AoG is growing quickly. The PCA was added for two reasons: people ask me about it all the time and the PCA folks make their data easy to find. ...

   "The mainline is just a bloodbath. Five traditions are down by at least 30%. The ELCA is down 41%. The United Church of Christ is less than half the size it was in the late 1980s. The United Methodists are already down 31%, but with over 15% of their churches disaffiliating just this year, I wouldn’t be surprised if membership is down 40% or more by this time next year.

   "The SBC is only down 4%, but that’s just because their decline has been way more recent." (Apparently, Burge goes "into a lot of detail" regarding the SBC numbers on graphsaboutreligion.com, but this appears to have been recently paywalled.)

   But two traditions "are up" - specifically, "The Assemblies of God has grown by over 50% in the past thirty-five years. The PCA has doubled in size, as well. But it’s important to put that PCA number in perspective. Even today, there are only about 400,000 members. For every PCA [member] in the United States there are about 33 Southern Baptists. ...

   The Assemblies of God's growth rate "has declined from 2% per annum to nearly 0% in the last few years. If that continues, the AoG might start shedding members, as well. ...

   Many "were already losing market share in the early 1990s. ...

   "For instance, the UCC was already way off the pace by 1995 and the gulf between the projection and the actual numbers just kept widening. For the ELCA it was a pretty small gap until about 2008, then it accelerated rapidly. That also happened for the SBC. Even into 2010, the difference between the lines is not huge but the last decade or so has been catastrophic.

   "Obviously the two outliers are the AoG and the PCA. They are growing faster than the natural increase in the population. The AoG growth is clearly less dramatic than the PCA, though. ...

   "If these nine denominations [had] grown as fast as the population from 1990 through 2022, they would have about 53 million members today. That’s about 16% of the overall population of the country.

   "Instead, they have a membership of 30.8 million people. That’s a total gap of just over 21 million members [who] don’t exist today. Those 30.8 million folks are just 9.3% of the total population. So, just these nine denominations are slightly more than half the size they should be if they continued to grow with America."

   Bottom line: "Denominations are in decline. Almost all of them, honestly. ...

   "Denominational Christianity used to be an incredibly important cultural force in American life. Leaders in these traditions used to hold sway over millions. Today, they are a shell of their former selves."

   Burge concludes: "The Southern Baptists are going through a bitter dispute that harkens back to the fights in the mid-1980s. The SBC emerged from those conflicts by continuing to enjoy membership growth. It seems unlikely that will be the case this time.

   "Meanwhile, the UMC will be drastically smaller next year than it was just two years ago in what can only be described as the largest denominational schism in the last fifty years.

   "These fights will not end in the near future. They will only accelerate. The big winner? That new non-denominational church down the road that has no institutional baggage." <www.bit.ly/3r89zDT>

   Before his paywall barrier went up, we had printed out this piece just to study Burge's graphs. The first, "The Decline in Membership of Nine Protestant Traditions," vividly shows that the AoG and PCA are the only groups with consistently positive growth. It also shows that the SBC has the steepest decline. Given the size of the SBC, one realizes that presumably, most of the overall Protestant denominational decline is comprised of Southern Baptists. It looks like this accounts for about two million of them since 2005.

   Sadly, in the graph "How Large Would Denominations Be If They Kept Pace with Population Growth?," we see that the SBC is also the group which *had* the steepest projected rise.

   A trend Burge missed: "Anglican Attendance Strongly Rebounds" (Juicy Ecumenism, Jun 6 '23) <www.tinyurl.com/43a2ab6m>

 ---

ARTIFICIAL INTELLIGENCE

When studying the sequence of explosive bursts <www.bit.ly/2KhtOIR> in technological history, we've wondered what the next "big thing" would be following the Internet's world-changing influence. What ever the marvel was, it seemed overdue. Surely it was already incubating. We've all heard AI is "big," and this essay certainly fuels such thinking. Is AI the "digital version" of the past era when the incremental production of specialized craftsmen morphed into "industrial manufacturing" mentioned below?

   If you think software coding is just for geeks, you sure won't like AI annotation. "AI Is a Lot of Work" by Josh Dzieza (New York Magazine's June cover story) is an informative inside look at the AI content preparation effort done by minions spread around the globe. At the front end of the process, each annotator in this "vast workforce mostly hidden behind the machines" ensures that the initial mountains of data are first sorted and tagged properly. This requires meticulous detail and pays only the lowest of wages. (No coding experience required.) The workers' united observation: "It's very boring." At the same time, it also reveals the ironic value of even the most basic applications of human intelligence.

   "As for the company employing them, most knew it only as Remotasks, a website offering work to anyone fluent in English. Like most of the annotators [Dzieza] spoke with, Joe was unaware until I told him that Remotasks is the worker-facing subsidiary of a company called Scale AI, a multibillion-dollar Silicon Valley data vendor that counts OpenAI and the U.S. military among its customers. ...

   "Much of the public response to language models like OpenAI's ChatGPT has focused on all the jobs they appear poised to automate. But behind even the most impressive AI system are people - huge numbers of people labeling data to train it and clarifying data when it gets confused."

   Most of us on the outside assume the so-called "machine learning" element of AI happens without human instruction. Dzieza explains this work isn't "employment without meaning or purpose, work that should be automated but for reasons of bureaucracy or status or inertia is not." Instead: "AI jobs are their bizarro twin: work that people want to automate, and often think is already automated, yet still requires a human stand-in. The jobs have a purpose; it's just that workers often have no idea what it is. ...

   "Mechanical Turk [is] Amazon's crowdsourcing platform where people around the world complete small tasks for cheap. The resulting annotated dataset, called ImageNet, enabled breakthroughs in machine learning that revitalized the field and ushered in a decade of progress. ...

   "Annotation remains a foundational part of making AI, but there is often a sense among engineers that it's a passing, inconvenient prerequisite [- and] annotation is never really finished. Machine-learning systems are what researchers call 'brittle,' prone to fail when encountering something that isn't well represented in their training data. These failures, called 'edge cases,' can have serious consequences."

   Currently, "there are no granular estimates of the number of people who work in annotation, but it is a lot, and it is growing. A recent Google Research paper gave an order-of-magnitude figure of 'millions' with the potential to become 'billions.'

   "AI doesn't replace work," Erik Duhaime, CEO of medical-data-annotation company Centaur Labs said. 'But it does change how work is organized.' ...

   "You might miss this if you believe AI is a brilliant, thinking machine. But if you pull back the curtain even a little, it looks more familiar, the latest iteration of a particularly Silicon Valley division of labor, in which the futuristic gleam of new technologies hides a sprawling manufacturing apparatus and the people who make it run. Duhaime reached back farther for a comparison, a digital version of the transition from craftsmen to industrial manufacturing: coherent processes broken into tasks and arrayed along assembly lines with some steps done by machines and some by humans but none resembling what came before. ...

   "When AI comes for your job, you may not lose it, but it might become more alien, more isolating, more tedious. ...

   "Where a human would get the concept of 'shirt' with a few examples, machine-learning programs need thousands, and they need to be categorized with perfect consistency yet varied enough (polo shirts, shirts being worn outdoors, shirts hanging on a rack) that the very literal system can handle the diversity of the real world. ...

   "When Remotasks first arrived in Kenya, annotators said it paid relatively well - averaging about $5 to $10 per hour depending on the task - but the amount fell as time went on. ...

   "Former Scale employees also said pay is determined through a surge-pricing-like mechanism that adjusts for how many annotators are available and how quickly the data is needed. ... 

   "By the beginning of this year, pay for the Kenyan annotators I spoke with had dropped to between $1 and $3 per hour. ...

   "[I]t's steady enough to be a full-time job for long stretches but too unpredictable to rely on. ...

   "A woman I'll call Anna was searching for a job in Texas when she stumbled across a generic listing for online work.... Her job is to talk with [a chatbot] all day. At about $14 an hour, plus bonuses for high productivity.... 

   "[T]he language that fuels ChatGPT and its competitors is filtered through several rounds of human annotation. ... After the model is trained on these examples, yet more contractors are brought in to prompt it and rank its responses. ... The point is that they are creating data on human taste, and once there's enough of it, engineers can train a second model to mimic their preferences at scale, automating the ranking process and training their AI to act in ways humans approve of. ...

   "Put another way, ChatGPT seems so human because it was trained by an AI that was mimicking humans who were rating an AI that was mimicking humans who were pretending to be a better version of an AI that was trained on human writing.

   "This circuitous technique is called 'reinforcement learning from human feedback,' or RLHF, and it's so effective that it's worth pausing to fully register what it doesn't do. ... There is no guarantee that the text the labelers marked as accurate is in fact accurate, and when it is, there is no guarantee that the model learns the right patterns from it.

   "This dynamic makes chatbot annotation a delicate process. ... 'Unlike many tasks in [machine learning] our queries do not have unambiguous ground truth,' [OpenAI researchers] lamented. ...

   "Because feedback data is difficult to collect, it fetches a higher price. ... Everyone involved is reluctant to say how much they're spending, but in general, specialized written examples can go for hundreds of dollars, while expert ratings can cost $50 or more. One engineer told me about buying examples of Socratic dialogues for up to $300 a pop."

   Dzieza includes speculation by leaders in the field on how some see annotation evolving and the challenges involved. "One way the AI industry differs from manufacturers of phones and cars is in its fluidity. The work is constantly changing, constantly getting automated away and replaced with new needs for new types of data. It's an assembly line but one that can be endlessly and instantly reconfigured, moving to wherever there is the right combination of skills, bandwidth, and wages."

   He closes with the example of how outclassed Kenyan annotators have started to game the system by masking their identity using a VPN (Virtual Private Network). The cat-and-mouse scramble continues on into the 21st century. <www.tinyurl.com/ym684556>

( - previous issue - / - next issue - )