Parrots, paper clips and safety vs. ethics: Why the artificial intelligence debate sounds like a foreign language
Sam Altman, chief executive officer and co-founder of OpenAI, speaks during a Senate Judiciary Subcommittee hearing in Washington, DC, US, on Tuesday, May 16, 2023. Congress is debating the potential and pitfalls of artificial intelligence as products like ChatGPT raise questions about the future of creative industries and the ability to tell fact from fiction.
Eric Lee | Bloomberg | Getty Images
This past week, OpenAI CEO Sam Altman charmed a room full of politicians in Washington, D.C., over dinner, then testified for about nearly three hours about potential risks of artificial intelligence at a Senate hearing.
After the hearing, he summed up his stance on AI regulation, using terms that are not widely known among the general public.
“AGI safety is really important, and frontier models should be regulated,” Altman tweeted. “Regulatory capture is bad, and we shouldn’t mess with models below the threshold.”
In this case, “AGI” refers to “artificial general intelligence.” As a concept, it’s used to mean a significantly more advanced AI than is currently possible, one that can do most things as well or better than most humans, including improving itself.
“Frontier models” is a way to talk about the AI systems that are the most expensive to produce and which analyze the most data. Large language models, like OpenAI’s GPT-4, are frontier models, as compared to smaller AI models that perform specific tasks like identifying cats in photos.
Most people agree that there need to be laws governing AI as the pace of development accelerates.
“Machine learning, deep learning, for the past 10 years or so, it developed very rapidly. When ChatGPT came out, it developed in a way we never imagined, that it could go this fast,” said My Thai, a computer science professor at the University of Florida. “We’re afraid that we’re racing into a more powerful system that we don’t fully comprehend and anticipate what what it is it can do.”
But the language around this debate reveals two major camps among academics, politicians, and the technology industry. Some are more concerned about what they call “AI safety.” The other camp is worried about what they call “AI ethics.“
When Altman spoke to Congress, he mostly avoided jargon, but his tweet suggested he’s mostly concerned about AI safety — a stance shared by many industry leaders at companies like Altman-run OpenAI, Google DeepMind and well-capitalized startups. They worry about the possibility of building an unfriendly AGI with unimaginable powers. This camp believes we need urgent attention from governments to regulate development an prevent an untimely end to humanity — an effort similar to nuclear nonproliferation.
“It’s good to hear so many people starting to get serious about AGI safety,” DeepMind founder and current Inflection AI CEO Mustafa Suleyman tweeted on Friday. “We need to be very ambitious. The Manhattan Project cost 0.4% of U.S. GDP. Imagine what an equivalent programme for safety could achieve today.”
But much of the discussion in Congress and at the White House about regulation is through an AI ethics lens, which focuses on current harms.
From this perspective, governments should enforce transparency around how AI systems collect and use data, restrict its use in areas that are subject to anti-discrimination law like housing or employment, and explain how current AI technology falls short. The White House’s AI Bill of Rights proposal from late last year included many of these concerns.
This camp was represented at the congressional hearing by IBM Chief Privacy Officer Christina Montgomery, who told lawmakers believes each company working on these technologies should have an “AI ethics” point of contact.
“There must be clear guidance on AI end uses or categories of AI-supported activity that are inherently high-risk,” Montgomery told Congress.
How to understand AI lingo like an insider
See also: How to talk about AI like an insider
It’s not surprising the debate around AI has developed its own lingo. It started as a technical academic field.
Much of the software being discussed today is based on so-called large language models (LLMs), which use graphic processing units (GPUs) to predict statistically likely sentences, images, or music, a process called “inference.” Of course, AI models need to be built first, in a data analysis process called “training.”
But other terms, especially from AI safety proponents, are more cultural in nature, and often refer to shared references and in-jokes.
For example, AI safety people might say that they’re worried about turning into a paper clip. That refers to a thought experiment popularized by philosopher Nick Bostrom that posits that a super-powerful AI — a “superintelligence” — could be given a mission to make as many paper clips as possible, and logically decide to kill humans make paper clips out of their remains.
OpenAI’s logo is inspired by this tale, and the company has even made paper clips in the shape of its logo.
Another concept in AI safety is the “hard takeoff” or “fast takeoff,” which is a phrase that suggests if someone succeeds at building an AGI that it will already be too late to save humanity.
Sometimes, this idea is described in terms of an onomatopeia — “foom” — especially among critics of the concept.
“It’s like you believe in the ridiculous hard take-off ‘foom’ scenario, which makes it sound like you have zero understanding of how everything works,” tweeted Meta AI chief Yann LeCun, who is skeptical of AGI claims, in a recent debate on social media.
AI ethics has its own lingo, too.
When describing the limitations of the current LLM systems, which cannot understand meaning but merely produce human-seeming language, AI ethics people often compare them to “Stochastic Parrots.“
The analogy, coined by Emily Bender, Timnit Gebru, Angelina McMillan-Major, and Margaret Mitchell in a paper written while some of the authors were at Google, emphasizes that while sophisticated AI models can produce realistic seeming text, the software doesn’t understand the concepts behind the language — like a parrot.
When these LLMs invent incorrect facts in responses, they’re “hallucinating.”
One topic IBM’s Montgomery pressed during the hearing was “explainability” in AI results. That means that when researchers and practitioners cannot point to the exact numbers and path of operations that larger AI models use to derive their output, this could hide some inherent biases in the LLMs.
“You have to have explainability around the algorithm,” said Adnan Masood, AI architect at UST-Global. “Previously, if you look at the classical algorithms, it tells you, ‘Why am I making that decision?’ Now with a larger model, they’re becoming this huge model, they’re a black box.”
Another important term is “guardrails,” which encompasses software and policies that Big Tech companies are currently building around AI models to ensure that they don’t leak data or produce disturbing content, which is often called “going off the rails.“
It can also refer to specific applications that protect AI software from going off topic, like Nvidia’s “NeMo Guardrails” product.
“Our AI ethics board plays a critical role in overseeing internal AI governance processes, creating reasonable guardrails to ensure we introduce technology into the world in a responsible and safe manner,” Montgomery said this week.
Sometimes these terms can have multiple meanings, as in the case of “emergent behavior.”
A recent paper from Microsoft Research called “sparks of artificial general intelligence” claimed to identify several “emergent behaviors” in OpenAI’s GPT-4, such as the ability to draw animals using a programming language for graphs.
But it can also describe what happens when simple changes are made at a very big scale — like the patterns birds make when flying in packs, or, in AI’s case, what happens when ChatGPT and similar products are being used by millions of people, such as widespread spam or disinformation.
Technology
UnitedHealth is making struggling doctors repay loans issued after last year’s cyberattack
UnitedHealth CEO Andrew Witty testifies before the Senate Finance Committee on Capitol Hill in Washington, May 1, 2024.
Kent Nishimura | Getty Images
Following the massive cyberattack on UnitedHealth Group’s Change Healthcare unit last year, the company launched a temporary funding assistance program to help medical practices with their short-term cash flow needs, offering no-interest loans with no added fees.
A little over a year later, UnitedHealth is aggressively going after borrowers, demanding they “immediately repay” their outstanding balances, according to documents viewed by CNBC and providers who received funding. Some groups have been asked to repay hundreds of thousands of dollars in a matter of days.
Optum, UnitedHealth’s financial, pharmacy and care services arm, is telling borrowers that it reserves the right to “begin offsetting claims payable” to the practices, meaning the company will withhold separate funds until it recoups the loan.
It’s a significant change in posture for the company, which suffered a cyberattack in February 2024 that compromised data from around 190 million Americans, the largest reported health-care breach in U.S. history. The ensuing disruption caused severe fallout across the health-care system, leaving many providers temporarily unable to get paid for their services. Some dipped into their personal savings to keep their practices afloat.
During a Senate hearing about the attack in May, UnitedHealth CEO Andrew Witty said providers would only be required to repay the loans when “they, not me, but they confirm that their cash flow is normalized.”
Several doctors who took advantage of the financing told CNBC that they can’t meet the company’s new demands. Dr. Christine Meyer, an internist who started a practice in Exton, Pennsylvania, received a letter from Optum earlier this month telling her to immediately submit her organization’s payment.
“We are not in any position to start repaying this loan,” Meyer, who started her practice about 20 years ago, told CNBC. She has been a vocal critic of UnitedHealth following the breach.
“I’m just looking at all my legal options at this point,” Meyer said. “But repaying them $750,000 in five days is obviously not going to happen.”
UnitedHealth didn’t comment on specific cases, but a spokesperson for Change Healthcare confirmed that the company has started recouping the loans.
“Now, more than one year post the event and with services restored, we have begun the process of recouping the interest-free funding we provided to providers,” the spokesperson said in a statement.
The company said the U.S. Department of Health and Human Services took the same approach last year “under its own cyber-attack lending program.” HHS launched a separate funding assistance program through the Centers for Medicare & Medicaid Services last March. CMS said it would automatically recoup payments from Medicare claims, and providers could accrue interest, according to a release.
“We continue to work with providers on repayment and other options, and continue to reach out to those providers that have not been responsive to previous calls or email requests for more information,” the Change Healthcare spokesperson said.
Providers were told that UnitedHealth reserved the right to withhold future payments when they signed up for the funding assistance program, the company added. CNBC independently reviewed a copy of a loan agreement for the program and confirmed this statement.
Change Healthcare, which offers payment and revenue cycle management tools, was acquired by Optum in 2022.
After discovering the breach last year, UnitedHealth said it isolated and disconnected the impacted systems. The company paid out more than $9 billion to providers in 2024, and more than $4.5 billion has already been repaid, according to the company’s fourth-quarter earnings report in January. UnitedHealth said providers would receive an invoice once standard payment operations resumed, and that they would be subject to a repayment period of 45 business days.
“Change Healthcare will notify the recipient that the funding amount is due after claims processing or payment processing services have resumed and payments impacted during the service disruption period are processed,” the website says.
Dwindling deposits, lost revenue
While the vast majority of Change Healthcare’s services have been restored over the course of the last year, three products are still listed as “partial service available,” according to UnitedHealth’s cyberattack response website.
And doctors are still reeling.
Meyer said that when the breach took place, she watched her practice’s daily deposits shrivel from the range of $60,000 to $80,000 to about $150 “overnight.” She applied for Optum’s temporary funding assistance program, and after some difficulty and back and forth with the company, she ultimately received a total of $756,900 in financial assistance.
Former Senator Bob Casey Jr., D-Pa., shared Meyer’s story during the congressional hearing in May. He asked Witty about the company’s approach to the repayment process.
“I’d like to absolutely confirm to you and Dr. Meyer that we have no intention of asking for loan repayment until after she determines that her business is back to normal,” Witty told lawmakers. “Even then, we would not look for repayment until 45 business days – 60 calendar days – after that and there would be no interest and no fee associated with that loan.”
“So it would be a determination she makes?” Casey asked.
“That’s absolutely right,” Witty said.
Meyer said that’s not what happened.
UnitedHealth Group Inc. headquarters stands in Minnetonka, Minnesota, U.S.
Mike Bradley | Bloomberg | Getty Images
She received a notice from Optum on Jan. 24, which was viewed by CNBC, that requested repayment since “the service disruption has ended for most clients.” Meyer said she called and told the company she was “not in any position to pay.”
Meyer claims that her practice lost more than $1 million in revenue due to the Change Healthcare cyberattack. She told CNBC the figure was based on a forensic financial analysis her practice carried out by comparing its charges against payments over recent years. The $1.2 million figure accounts for losses across all its insurers, not just UnitedHealthcare, Meyer said.
On April 1, Meyer received another notice requesting immediate repayment within five business days. The letter was addressed to Meyer. But the name of the practice on the letter, Insight Counseling, as well as the total amount due, $925,200, were incorrect.
Meyer said she called Optum again and was told the company made a mistake, but that she had five days to repay her actual total of $750,000. At that point, the company would start withholding her UnitedHealthcare payments, which she described as a “shakedown.”
Meyer said her practice typically receives annual claims payments of about $150,000 to $200,000 from UnitedHealthcare.
“I guess I’ll just let them take those payments back for the next three years until they get their money back,” she told CNBC.
In a post on LinkedIn on Thursday, Meyer wrote that she and her team “made a plan to leave the least amount of money in the account set up to receive payments from UnitedHealthcare. If it isn’t there, they can’t get it.”
‘Very frustrating experience’
Dr. Purvi Parikh, an allergist and immunologist with a private practice in New York, shared a similar story.
Parikh’s practice received about $440,000 in funding assistance after the breach. She said she started getting repayment notices late last year, and that Optum was threatening to offset claims payable to the practice.
“We were already hit very hard by the Change Healthcare hack,” Parikh said in an interview. “Now on top of that, they’re asking for all of this money back or they’re going to hold future payments ransom. It’s just been a very frustrating experience dealing with Optum.”
Parikh’s practice requested a one-month extension on its final payment of $101,650 in January to try and keep UnitedHealth from withholding other payments. In the email request, Parikh’s colleague wrote that “it has been quite difficult to recover financially.”
Optum granted Parikh’s practice the extension.
“People don’t just have that amount of money just sitting around,” Parikh said. “We’ve paid everything back, but it wasn’t without hardship.”
A physician who runs a pediatric practice in New Jersey said UnitedHealth has already started withholding payments from the organization. The practice received more than $500,000 in funding assistance following the Change Healthcare breach.
The doctor, who asked not to be named due to the sensitive nature of the situation, said the practice began receiving phone calls and emails from Optum requesting repayment beginning late last year. The group indicated that it didn’t have the money, but would set up a payment plan and had begun the process.
But the doctor said its billing department noticed that UnitedHealth had already started holding back claims payments. In its explanation of benefits, which details what an insurer will cover, the doctor said the company has a line that reads, “UnitedHealthcare is withholding payment for Optum.”
WATCH: Health and Human Services Department opens probe into hack at UnitedHealth’s Change Healthcare
-
Sports2 years ago‘Storybook stuff’: Inside the night Bryce Harper sent the Phillies to the World Series
-
Sports1 year agoStory injured on diving stop, exits Red Sox game
-
Sports1 year agoGame 1 of WS least-watched in recorded history
-
Sports2 years agoMLB Rank 2023: Ranking baseball’s top 100 players
-
Sports4 years ago
Team Europe easily wins 4th straight Laver Cup
-
Environment2 years agoJapan and South Korea have a lot at stake in a free and open South China Sea
-
Environment2 years agoGame-changing Lectric XPedition launched as affordable electric cargo bike
-
Business3 years agoBank of England’s extraordinary response to government policy is almost unthinkable | Ed Conway