Connect with us

Spread & Containment

Who Are the Federal Student Loan Borrowers and Who Benefits from Forgiveness?

The pandemic forbearance for federal student loans was recently extended for a sixth time—marking a historic thirty-month pause on federal student loan…

Published

on

The pandemic forbearance for federal student loans was recently extended for a sixth time—marking a historic thirty-month pause on federal student loan payments. The first post in this series uses survey data to help us understand which borrowers are likely to struggle when the pandemic forbearance ends. The results from this survey and the experience of some federal borrowers who did not receive forbearance during the pandemic suggest that delinquencies could surpass pre-pandemic levels after forbearance ends. These concerns have revived debates over the possibility of blanket forgiveness of federal student loans. Calls for student loan forgiveness entered the mainstream during the 2020 election with most proposals centering around blanket federal student loan forgiveness (typically $10,000 or $50,000) or loan forgiveness with certain income limits for eligibility. Several studies (examples here, here, and here) have attempted to quantify the costs and distribution of benefits of some of these policies. However, each of these studies either relies on data that do not fully capture the population that owes student loan debt or does not separate student loans owned by the federal government from those owned by commercial banks and are thus not eligible for forgiveness with most proposals. In this post, we use representative data from anonymized credit reports that allows us to identify federal loans, calculate the total cost of these proposals, explore important heterogeneity in who owes federal student loans, and examine who would likely benefit from federal student loan forgiveness.

This analysis offers a granular look at outcomes under various policy options. We find that smaller forgiveness policies distribute a greater share of benefit to borrowers with low- and mid-range credit scores and residing in low- and middle-income neighborhoods. Increasing the per-borrower maximum forgiveness shifts larger shares of forgiven debt to higher credit score borrowers and higher income neighborhoods. By contrast, limiting forgiveness eligibility by income reduces the total cost of the policy while distributing larger shares of forgiveness to low- and middle-income neighborhoods, low- and mid-credit score borrowers, and majority minority neighborhoods.  

Data and Definitions

We use the New York Fed/Equifax Consumer Credit Panel (CCP) which is a nationally representative 5 percent sample of all U.S. adults with a credit report. We directly observe a borrower’s age, credit score, and student loan balance, but we do not observe an individual’s income or demographic information. Instead, we use Census block group identifiers from the CCP to match an individual to information about their neighborhood, such as median household income and demographics, from the five-year American Community Survey 2014-2018. We identify student loans that are held by the federal government by selecting loans that entered automatic administrative forbearance at the beginning of the COVID-19 pandemic. These include Direct loans that were disbursed by the federal government and loans originally disbursed through the Family Federal Education Loan (FFEL) Program but were subsequently consolidated into the Direct program or sold to the federal government. These also include loans disbursed from either the Direct or FFEL program that are in default.

Costs of Forgiveness Policies

We estimate the total cost of federal loan forgiveness policies by calculating the dollar value of the loans that would be forgiven under each policy. We limit the sample of loans eligible for forgiveness to only those owned by the federal government since this has been the focus of most cancellation proposals. The total outstanding balance for federally-owned (including defaulted) student loans in December 2021 was $1.38 trillion. Limiting forgiveness to a maximum of $50,000 per borrower would cost $904 billion and would forgive the full balance for 29.9 million (79 percent) of the 37.9 million federal borrowers, resulting in an average forgiveness of $23,856 per borrower. This threshold would also forgive 77 percent of all federal student loans that were delinquent or in default prior to the pandemic. Meanwhile, forgiveness of $10,000 per borrower would forgive a total of $321 billion of federal student loans, eliminate the entire balance for 11.8 million borrowers (31.1 percent), and cancel 30.5 percent of loans delinquent or in default prior to the pandemic forbearance. Under this policy, the average borrower would receive $8,478 in student loan forgiveness.

Next, we explore the impact of adding income limits for determining eligibility for forgiveness. Since we do not directly observe a borrower’s income, we simulate eligibility by sampling from the distribution of household income for each borrower’s Census block group and take the average total forgiveness over 100 simulations. Adding a household income limit of $75,000 reduces the total cost of a $50,000 forgiveness policy from $904 billion to $507 billion, a reduction of almost 45 percent. Similarly, the same income limit reduces the cost of a $10,000 forgiveness policy from $321 billion to $182 billion.

One caveat is that the estimate for the cost of potential student loan forgiveness policies is likely the upper bound. Specifically, some of the balances forgiven under these hypothetical blanket policies will eventually be forgiven under the Public Service Loan Forgiveness (PSLF) program or through income-driven repayment plans. For these loans, the net cost of blanket forgiveness now would not be the total outstanding amount of each loan (as we calculate) but instead would be the stream of monthly payments on these loans until they are cancelled under existing forgiveness policies.

Who Benefits from Forgiveness?

BY AGE

We begin by studying who holds federal student loan balances and who would receive forgiveness by age under the various policies. Sixty-seven percent of student loan borrowers are under 40, however only 57 percent of balances are owed by those under 40, showing that those with larger balances are more likely to be older (likely due to borrowing for graduate school). Under each of the considered policies (forgiveness at the $10,000 or the $50,000 level, with and without income caps), over 60 percent of forgiven loan dollars benefit those under 40 years old. While income caps do not significantly change the share of forgiveness going to each age group, increasing the forgiveness amount from $10,000 to $50,000 shifts a larger share of forgiven debt to older borrowers. However, those over 60 years old benefit the least from forgiveness. Despite being 32 percent of the U.S. adult population, those 60 and older only receive around 6 percent of forgiven dollars, roughly in line with the share of this age group that owes federal student loans.

Student Loan Forgiveness Predominately Benefits Those Under 40

Sources: New York Fed/Equifax Consumer Credit Panel; authors’ calculations.
Note: Total shares for each policy may not sum to 100 percent due to rounding or missing identifiers.

By Neighborhood Income

Next, we study who benefits from student loan forgiveness by income. Since we do not directly observe income for individuals in the data, we assign individuals to an income category by the median income of their neighborhood through Census block group designations. We split income into quartiles with the lowest quartile defined as low-income (with a median annual income below $46,310), the middle two quartiles as middle-income (between $46,310 and $78,303 per year), and the highest quartile as high-income ($78,303 and above per year). Borrowers living in high-income areas are more likely to owe federal student loans and hold higher balances. Despite being 25 percent of the population, borrowers who live in high-income neighborhoods hold 33 percent of federal balances while borrowers residing in low-income areas hold only 23 percent of balances. Under both forgiveness levels without income caps, low-income neighborhoods receive roughly 25 percent of debt forgiveness while high-income neighborhoods receive around 30 percent of forgiveness. Increasing the threshold from $10,000 to $50,000 results in a marginally larger share of forgiveness to high-income areas. The average federal student loan borrower living in a high-income neighborhood would receive $25,054 while the average borrower living in a low-income neighborhood would receive $22,512. By contrast, adding a $75,000 income cap for forgiveness eligibility significantly shifts the share of benefits. The share of forgiven dollars going to high-income areas falls from around 30 percent to around 18 percent and the share of forgiven debt going to low-income areas increases from around 25 percent to around 34 percent.

Increasing Student Loan Forgiveness Distributes a Larger Share of Benefits to Higher-Income Neighborhoods, but Income Caps Counterbalance this Trend

Sources: New York Fed/Equifax Consumer Credit Panel; American Community Survey 2014-2018; authors’ calculations.
Notes: We assign individuals to an income category by the median income of their neighborhood through Census block group designations. The low-income group represents those with a neighborhood income median below $46,310 per year, the middle-income group between $46,310 and $78,303, and the high-income group $78,303 or more. Total shares for each policy may not sum to 100 percent due to rounding or missing identifiers.

By Credit Score

We also track the share of federal student loan forgiveness that would benefit people with different levels of financial stability by categorizing them into credit score bins. Credit scores serve as a proxy for both income and financial stability so borrowers with lower credit scores are more likely to struggle with payments while borrowers with higher scores are more likely to be of higher income and more financially stable. We use credit scores from February 2020 since previously delinquent federal student loan borrowers experienced large credit score increases when their accounts were marked current due to pandemic forbearance. Compared to the population of U.S. adults with a credit report, student loan borrowers have substantially lower credit scores. Roughly 34 percent of all credit scores are greater than 760, but only 11 percent of student loan borrowers have these super prime scores. When weighted by balance, student loan borrowers have higher scores suggesting that those with high balances also have higher credit scores. Under all four policies, more than half the share of forgiven debt would go to borrowers with a credit score below 660. As with our analysis by income, increasing the threshold from $10,000 to $50,000 increases the share of forgiven balances going to those with credit scores of 720 or higher, suggesting that a higher per borrower forgiveness amount tends to benefit borrowers of higher socioeconomic status more. However, income caps reduce the share of benefits going to those with super prime scores and distributes a larger share of forgiveness to those with lower credit scores.

Most Forgiven Debt Would Go to Student Loan Borrowers with Lower Credit Scores

Sources: New York Fed/Equifax Consumer Credit Panel; authors’ calculations.
Note: Total shares for each policy may not sum to 100 percent due to rounding or missing identifiers.

By Neighborhood Demographics

We next examine who benefits from forgiveness based on demographic characteristics of a borrower’s neighborhood. We separate borrowers into two categories: those who live in a Census block group with more than 50 percent of residents identifying as white non-Hispanic (majority white) and those who live in a Census block group with at most 50 percent white non-Hispanic residents (majority minority). Those living in majority white and majority minority neighborhoods are equally likely to owe student loans; roughly 67 percent of the population and 67 percent of federal student loan borrowers reside in majority white neighborhoods and balance shares are split roughly in the same proportion. Under a $10,000 forgiveness policy, 33 percent of forgiveness would go to majority minority neighborhoods while 67 percent would go to majority white neighborhoods. Further increasing forgiveness from $10,000 to $50,000 does not significantly change these shares. However, introducing an income cap of $75,000 for eligibility significantly increases the share of forgiven loans going to majority minority neighborhoods—from roughly 33 percent of forgiven debt to 37 percent at both forgiveness levels.

Income Caps Shift a Larger Share of Forgiven Student Loans to Majority Minority Neighborhoods

Sources: New York Fed/Equifax Consumer Credit Panel; American Community Survey 2014-2018; authors’ calculations.
Notes: We separate borrowers into two categories: those who live in a Census block group with at most 50 percent white non-Hispanic residents (majority minority) and those who live in a Census block group with more than 50 percent of residents identifying as white non-Hispanic (majority white). Total shares for each policy may not sum to 100 percent due to rounding or missing identifiers.

Conclusion

In this post, we examine who benefits from various federal student loan forgiveness proposals. In general, we find that smaller student loan forgiveness policies distribute a larger share of benefits to lower credit score borrowers and to those that live in less wealthy and majority minority neighborhoods (relative to the share of balances they hold). Increasing the forgiveness amount increases the share of total forgiven debt for higher credit score borrowers and those living in richer neighborhoods with a majority of white residents.

We find that adding an income cap to forgiveness proposals substantially reduces the cost of student loan forgiveness and increases the share of benefit going to borrowers who are more likely to struggle repaying their debts. A $75,000 income cap drops the cost of forgiveness by almost 45 percent for either a $10,000 or $50,000 policy. Further, it drastically changes the distribution of benefits. Under a $10,000 policy, an income cap raises the share of forgiven loan dollars going to borrowers in low-income neighborhoods from 25 percent to 35 percent and the share going to lower credit score borrowers from 37 percent to 42 percent. Income caps also increase the share of loans forgiven that were delinquent prior to the pandemic. Adding an income cap to a $10,000 policy increases the share of forgiveness canceling loans that were delinquent before the pandemic from 34 percent to 60 percent. Under any policy, means testing would more directly target forgiveness to borrowers facing a greater struggle with repayment, which would result in a significantly less regressive policy.

Jacob Goss is a senior research analyst in the Federal Reserve Bank of New York’s Research and Statistics Group.

Daniel Mangrum is an economist in the Bank’s Research and Statistics Group.

Joelle Scally is a senior data strategist in the Bank’s Research and Statistics Group.

How to cite this post:
Jacob Goss, Daniel Mangrum, and Joelle Scally, “Who Are the Federal Student Loan Borrowers and Who Benefits from Forgiveness?,” Federal Reserve Bank of New York Liberty Street Economics, April 21, 2022, https://libertystreeteconomics.newyorkfed.org/who-are-the-federal-student-loan-borrowers-and-who-benefits-from-forgiveness/.


Disclaimer
The views expressed in this post are those of the authors and do not necessarily reflect the position of the Federal Reserve Bank of New York or the Federal Reserve System. Any errors or omissions are the responsibility of the authors.

Disclosure
An author owes federal student loans that could be cancelled under some of these policies.

Read More

Continue Reading

Government

“I Can’t Even Save”: Americans Are Getting Absolutely Crushed Under Enormous Debt Load

"I Can’t Even Save": Americans Are Getting Absolutely Crushed Under Enormous Debt Load

While Joe Biden insists that Americans are doing great…

Published

on

"I Can't Even Save": Americans Are Getting Absolutely Crushed Under Enormous Debt Load

While Joe Biden insists that Americans are doing great - suggesting in his State of the Union Address last week that "our economy is the envy of the world," Americans are being absolutely crushed by inflation (which the Biden admin blames on 'shrinkflation' and 'corporate greed'), and of course - crippling debt.

The signs are obvious. Last week we noted that banks' charge-offs are accelerating, and are now above pre-pandemic levels.

...and leading this increase are credit card loans - with delinquencies that haven't been this high since Q3 2011.

On top of that, while credit cards and nonfarm, nonresidential commercial real estate loans drove the quarterly increase in the noncurrent rate, residential mortgages drove the quarterly increase in the share of loans 30-89 days past due.

And while Biden and crew can spin all they want, an average of polls from RealClear Politics shows that just 40% of people approve of Biden's handling of the economy.

Crushed

On Friday, Bloomberg dug deeper into the effects of Biden's "envious" economy on Americans - specifically, how massive debt loads (credit cards and auto loans especially) are absolutely crushing people.

Two years after the Federal Reserve began hiking interest rates to tame prices, delinquency rates on credit cards and auto loans are the highest in more than a decade. For the first time on record, interest payments on those and other non-mortgage debts are as big a financial burden for US households as mortgage interest payments.

According to the report, this presents a difficult reality for millions of consumers who drive the US economy - "The era of high borrowing costs — however necessary to slow price increases — has a sting of its own that many families may feel for years to come, especially the ones that haven’t locked in cheap home loans."

The Fed, meanwhile, doesn't appear poised to cut rates until later this year.

According to a February paper from IMF and Harvard, the recent high cost of borrowing - something which isn't reflected in inflation figures, is at the heart of lackluster consumer sentiment despite inflation having moderated and a job market which has recovered (thanks to job gains almost entirely enjoyed by immigrants).

In short, the debt burden has made life under President Biden a constant struggle throughout America.

"I’m making the most money I've ever made, and I’m still living paycheck to paycheck," 40-year-old Denver resident Nikki Cimino told Bloomberg. Cimino is carrying a monthly mortgage of $1,650, and has $4,000 in credit card debt following a 2020 divorce.

Nikki CiminoPhotographer: Rachel Woolf/Bloomberg

"There's this wild disconnect between what people are experiencing and what economists are experiencing."

What's more, according to Wells Fargo, families have taken on debt at a comparatively fast rate - no doubt to sustain the same lifestyle as low rates and pandemic-era stimmies provided. In fact, it only took four years for households to set a record new debt level after paying down borrowings in 2021 when interest rates were near zero. 

Meanwhile, that increased debt load is exacerbated by credit card interest rates that have climbed to a record 22%, according to the Fed.

[P]art of the reason some Americans were able to take on a substantial load of non-mortgage debt is because they’d locked in home loans at ultra-low rates, leaving room on their balance sheets for other types of borrowing. The effective rate of interest on US mortgage debt was just 3.8% at the end of last year.

Yet the loans and interest payments can be a significant strain that shapes families’ spending choices. -Bloomberg

And of course, the highest-interest debt (credit cards) is hurting lower-income households the most, as tends to be the case.

The lowest earners also understandably had the biggest increase in credit card delinquencies.

"Many consumers are levered to the hilt — maxed out on debt and barely keeping their heads above water," Allan Schweitzer, a portfolio manager at credit-focused investment firm Beach Point Capital Management told Bloomberg. "They can dog paddle, if you will, but any uptick in unemployment or worsening of the economy could drive a pretty significant spike in defaults."

"We had more money when Trump was president," said Denise Nierzwicki, 69. She and her 72-year-old husband Paul have around $20,000 in debt spread across multiple cards - all of which have interest rates above 20%.

Denise and Paul Nierzwicki blame Biden for what they see as a gloomy economy and plan to vote for the Republican candidate in November.
Photographer: Jon Cherry/Bloomberg

During the pandemic, Denise lost her job and a business deal for a bar they owned in their hometown of Lexington, Kentucky. While they applied for Social Security to ease the pain, Denise is now working 50 hours a week at a restaurant. Despite this, they're barely scraping enough money together to service their debt.

The couple blames Biden for what they see as a gloomy economy and plans to vote for the Republican candidate in November. Denise routinely voted for Democrats up until about 2010, when she grew dissatisfied with Barack Obama’s economic stances, she said. Now, she supports Donald Trump because he lowered taxes and because of his policies on immigration. -Bloomberg

Meanwhile there's student loans - which are not able to be discharged in bankruptcy.

"I can't even save, I don't have a savings account," said 29-year-old in Columbus, Ohio resident Brittany Walling - who has around $80,000 in federal student loans, $20,000 in private debt from her undergraduate and graduate degrees, and $6,000 in credit card debt she accumulated over a six-month stretch in 2022 while she was unemployed.

"I just know that a lot of people are struggling, and things need to change," she told the outlet.

The only silver lining of note, according to Bloomberg, is that broad wage gains resulting in large paychecks has made it easier for people to throw money at credit card bills.

Yet, according to Wells Fargo economist Shannon Grein, "As rates rose in 2023, we avoided a slowdown due to spending that was very much tied to easy access to credit ... Now, credit has become harder to come by and more expensive."

According to Grein, the change has posed "a significant headwind to consumption."

Then there's the election

"Maybe the Fed is done hiking, but as long as rates stay on hold, you still have a passive tightening effect flowing down to the consumer and being exerted on the economy," she continued. "Those household dynamics are going to be a factor in the election this year."

Meanwhile, swing-state voters in a February Bloomberg/Morning Consult poll said they trust Trump more than Biden on interest rates and personal debt.

Reverberations

These 'headwinds' have M3 Partners' Moshin Meghji concerned.

"Any tightening there immediately hits the top line of companies," he said, noting that for heavily indebted companies that took on debt during years of easy borrowing, "there's no easy fix."

Tyler Durden Fri, 03/15/2024 - 18:00

Read More

Continue Reading

Spread & Containment

Sylvester researchers, collaborators call for greater investment in bereavement care

MIAMI, FLORIDA (March 15, 2024) – The public health toll from bereavement is well-documented in the medical literature, with bereaved persons at greater…

Published

on

MIAMI, FLORIDA (March 15, 2024) – The public health toll from bereavement is well-documented in the medical literature, with bereaved persons at greater risk for many adverse outcomes, including mental health challenges, decreased quality of life, health care neglect, cancer, heart disease, suicide, and death. Now, in a paper published in The Lancet Public Health, researchers sound a clarion call for greater investment, at both the community and institutional level, in establishing support for grief-related suffering.

Credit: Photo courtesy of Memorial Sloan Kettering Comprehensive Cancer Center

MIAMI, FLORIDA (March 15, 2024) – The public health toll from bereavement is well-documented in the medical literature, with bereaved persons at greater risk for many adverse outcomes, including mental health challenges, decreased quality of life, health care neglect, cancer, heart disease, suicide, and death. Now, in a paper published in The Lancet Public Health, researchers sound a clarion call for greater investment, at both the community and institutional level, in establishing support for grief-related suffering.

The authors emphasized that increased mortality worldwide caused by the COVID-19 pandemic, suicide, drug overdose, homicide, armed conflict, and terrorism have accelerated the urgency for national- and global-level frameworks to strengthen the provision of sustainable and accessible bereavement care. Unfortunately, current national and global investment in bereavement support services is woefully inadequate to address this growing public health crisis, said researchers with Sylvester Comprehensive Cancer Center at the University of Miami Miller School of Medicine and collaborating organizations.  

They proposed a model for transitional care that involves firmly establishing bereavement support services within healthcare organizations to ensure continuity of family-centered care while bolstering community-based support through development of “compassionate communities” and a grief-informed workforce. The model highlights the responsibility of the health system to build bridges to the community that can help grievers feel held as they transition.   

The Center for the Advancement of Bereavement Care at Sylvester is advocating for precisely this model of transitional care. Wendy G. Lichtenthal, PhD, FT, FAPOS, who is Founding Director of the new Center and associate professor of public health sciences at the Miller School, noted, “We need a paradigm shift in how healthcare professionals, institutions, and systems view bereavement care. Sylvester is leading the way by investing in the establishment of this Center, which is the first to focus on bringing the transitional bereavement care model to life.”

What further distinguishes the Center is its roots in bereavement science, advancing care approaches that are both grounded in research and community-engaged.  

The authors focused on palliative care, which strives to provide a holistic approach to minimize suffering for seriously ill patients and their families, as one area where improvements are critically needed. They referenced groundbreaking reports of the Lancet Commissions on the value of global access to palliative care and pain relief that highlighted the “undeniable need for improved bereavement care delivery infrastructure.” One of those reports acknowledged that bereavement has been overlooked and called for reprioritizing social determinants of death, dying, and grief.

“Palliative care should culminate with bereavement care, both in theory and in practice,” explained Lichtenthal, who is the article’s corresponding author. “Yet, bereavement care often is under-resourced and beset with access inequities.”

Transitional bereavement care model

So, how do health systems and communities prioritize bereavement services to ensure that no bereaved individual goes without needed support? The transitional bereavement care model offers a roadmap.

“We must reposition bereavement care from an afterthought to a public health priority. Transitional bereavement care is necessary to bridge the gap in offerings between healthcare organizations and community-based bereavement services,” Lichtenthal said. “Our model calls for health systems to shore up the quality and availability of their offerings, but also recognizes that resources for bereavement care within a given healthcare institution are finite, emphasizing the need to help build communities’ capacity to support grievers.”

Key to the model, she added, is the bolstering of community-based support through development of “compassionate communities” and “upskilling” of professional services to assist those with more substantial bereavement-support needs.

The model contains these pillars:

  • Preventive bereavement care –healthcare teams engage in bereavement-conscious practices, and compassionate communities are mindful of the emotional and practical needs of dying patients’ families.
  • Ownership of bereavement care – institutions provide bereavement education for staff, risk screenings for families, outreach and counseling or grief support. Communities establish bereavement centers and “champions” to provide bereavement care at workplaces, schools, places of worship or care facilities.
  • Resource allocation for bereavement care – dedicated personnel offer universal outreach, and bereaved stakeholders provide input to identify community barriers and needed resources.
  • Upskilling of support providers – Bereavement education is integrated into training programs for health professionals, and institutions offer dedicated grief specialists. Communities have trained, accessible bereavement specialists who provide support and are educated in how to best support bereaved individuals, increasing their grief literacy.
  • Evidence-based care – bereavement care is evidence-based and features effective grief assessments, interventions, and training programs. Compassionate communities remain mindful of bereavement care needs.

Lichtenthal said the new Center will strive to materialize these pillars and aims to serve as a global model for other health organizations. She hopes the paper’s recommendations “will cultivate a bereavement-conscious and grief-informed workforce as well as grief-literate, compassionate communities and health systems that prioritize bereavement as a vital part of ethical healthcare.”

“This paper is calling for healthcare institutions to respond to their duty to care for the family beyond patients’ deaths. By investing in the creation of the Center for the Advancement of Bereavement Care, Sylvester is answering this call,” Lichtenthal said.

Follow @SylvesterCancer on X for the latest news on Sylvester’s research and care.

# # #

Article Title: Investing in bereavement care as a public health priority

DOI: 10.1016/S2468-2667(24)00030-6

Authors: The complete list of authors is included in the paper.

Funding: The authors received funding from the National Cancer Institute (P30 CA240139 Nimer) and P30 CA008748 Vickers).

Disclosures: The authors declared no competing interests.

# # #


Read More

Continue Reading

Spread & Containment

Separating Information From Disinformation: Threats From The AI Revolution

Separating Information From Disinformation: Threats From The AI Revolution

Authored by Per Bylund via The Mises Institute,

Artificial intelligence…

Published

on

Separating Information From Disinformation: Threats From The AI Revolution

Authored by Per Bylund via The Mises Institute,

Artificial intelligence (AI) cannot distinguish fact from fiction. It also isn’t creative or can create novel content but repeats, repackages, and reformulates what has already been said (but perhaps in new ways).

I am sure someone will disagree with the latter, perhaps pointing to the fact that AI can clearly generate, for example, new songs and lyrics. I agree with this, but it misses the point. AI produces a “new” song lyric only by drawing from the data of previous song lyrics and then uses that information (the inductively uncovered patterns in it) to generate what to us appears to be a new song (and may very well be one). However, there is no artistry in it, no creativity. It’s only a structural rehashing of what exists.

Of course, we can debate to what extent humans can think truly novel thoughts and whether human learning may be based solely or primarily on mimicry. However, even if we would—for the sake of argument—agree that all we know and do is mere reproduction, humans have limited capacity to remember exactly and will make errors. We also fill in gaps with what subjectively (not objectively) makes sense to us (Rorschach test, anyone?). Even in this very limited scenario, which I disagree with, humans generate novelty beyond what AI is able to do.

Both the inability to distinguish fact from fiction and the inductive tether to existent data patterns are problems that can be alleviated programmatically—but are open for manipulation.

Manipulation and Propaganda

When Google launched its Gemini AI in February, it immediately became clear that the AI had a woke agenda. Among other things, the AI pushed woke diversity ideals into every conceivable response and, among other things, refused to show images of white people (including when asked to produce images of the Founding Fathers).

Tech guru and Silicon Valley investor Marc Andreessen summarized it on X (formerly Twitter): “I know it’s hard to believe, but Big Tech AI generates the output it does because it is precisely executing the specific ideological, radical, biased agenda of its creators. The apparently bizarre output is 100% intended. It is working as designed.”

There is indeed a design to these AIs beyond the basic categorization and generation engines. The responses are not perfectly inductive or generative. In part, this is necessary in order to make the AI useful: filters and rules are applied to make sure that the responses that the AI generates are appropriate, fit with user expectations, and are accurate and respectful. Given the legal situation, creators of AI must also make sure that the AI does not, for example, violate intellectual property laws or engage in hate speech. AI is also designed (directed) so that it does not go haywire or offend its users (remember Tay?).

However, because such filters are applied and the “behavior” of the AI is already directed, it is easy to take it a little further. After all, when is a response too offensive versus offensive but within the limits of allowable discourse? It is a fine and difficult line that must be specified programmatically.

It also opens the possibility for steering the generated responses beyond mere quality assurance. With filters already in place, it is easy to make the AI make statements of a specific type or that nudges the user in a certain direction (in terms of selected facts, interpretations, and worldviews). It can also be used to give the AI an agenda, as Andreessen suggests, such as making it relentlessly woke.

Thus, AI can be used as an effective propaganda tool, which both the corporations creating them and the governments and agencies regulating them have recognized.

Misinformation and Error

States have long refused to admit that they benefit from and use propaganda to steer and control their subjects. This is in part because they want to maintain a veneer of legitimacy as democratic governments that govern based on (rather than shape) people’s opinions. Propaganda has a bad ring to it; it’s a means of control.

However, the state’s enemies—both domestic and foreign—are said to understand the power of propaganda and do not hesitate to use it to cause chaos in our otherwise untainted democratic society. The government must save us from such manipulation, they claim. Of course, rarely does it stop at mere defense. We saw this clearly during the covid pandemic, in which the government together with social media companies in effect outlawed expressing opinions that were not the official line (see Murthy v. Missouri).

AI is just as easy to manipulate for propaganda purposes as social media algorithms but with the added bonus that it isn’t only people’s opinions and that users tend to trust that what the AI reports is true. As we saw in the previous article on the AI revolution, this is not a valid assumption, but it is nevertheless a widely held view.

If the AI then can be instructed to not comment on certain things that the creators (or regulators) do not want people to see or learn, then it is effectively “memory holed.” This type of “unwanted” information will not spread as people will not be exposed to it—such as showing only diverse representations of the Founding Fathers (as Google’s Gemini) or presenting, for example, only Keynesian macroeconomic truths to make it appear like there is no other perspective. People don’t know what they don’t know.

Of course, nothing is to say that what is presented to the user is true. In fact, the AI itself cannot distinguish fact from truth but only generates responses according to direction and only based on whatever the AI has been fed. This leaves plenty of scope for the misrepresentation of the truth and can make the world believe outright lies. AI, therefore, can easily be used to impose control, whether it is upon a state, the subjects under its rule, or even a foreign power.

The Real Threat of AI

What, then, is the real threat of AI? As we saw in the first article, large language models will not (cannot) evolve into artificial general intelligence as there is nothing about inductive sifting through large troves of (humanly) created information that will give rise to consciousness. To be frank, we haven’t even figured out what consciousness is, so to think that we will create it (or that it will somehow emerge from algorithms discovering statistical language correlations in existing texts) is quite hyperbolic. Artificial general intelligence is still hypothetical.

As we saw in the second article, there is also no economic threat from AI. It will not make humans economically superfluous and cause mass unemployment. AI is productive capital, which therefore has value to the extent that it serves consumers by contributing to the satisfaction of their wants. Misused AI is as valuable as a misused factory—it will tend to its scrap value. However, this doesn’t mean that AI will have no impact on the economy. It will, and already has, but it is not as big in the short-term as some fear, and it is likely bigger in the long-term than we expect.

No, the real threat is AI’s impact on information. This is in part because induction is an inappropriate source of knowledge—truth and fact are not a matter of frequency or statistical probabilities. The evidence and theories of Nicolaus Copernicus and Galileo Galilei would get weeded out as improbable (false) by an AI trained on all the (best and brightest) writings on geocentrism at the time. There is no progress and no learning of new truths if we trust only historical theories and presentations of fact.

However, this problem can probably be overcome by clever programming (meaning implementing rules—and fact-based limitations—to the induction problem), at least to some extent. The greater problem is the corruption of what AI presents: the misinformation, disinformation, and malinformation that its creators and administrators, as well as governments and pressure groups, direct it to create as a means of controlling or steering public opinion or knowledge.

This is the real danger that the now-famous open letter, signed by Elon Musk, Steve Wozniak, and others, pointed to:

“Should we let machines flood our information channels with propaganda and untruth? Should we automate away all the jobs, including the fulfilling ones? Should we develop nonhuman minds that might eventually outnumber, outsmart, obsolete and replace us? Should we risk loss of control of our civilization?”

Other than the economically illiterate reference to “automat[ing] away all the jobs,” the warning is well-taken. AI will not Terminator-like start to hate us and attempt to exterminate mankind. It will not make us all into biological batteries, as in The Matrix. However, it will—especially when corrupted—misinform and mislead us, create chaos, and potentially make our lives “solitary, poor, nasty, brutish and short.”

Tyler Durden Fri, 03/15/2024 - 06:30

Read More

Continue Reading

Trending