ChatGPT -- Smart? Stupid? Benign? Dangerous?

A face for ChatGPT?
(Image by Robert Adler/DALL-E)   Details   DMCA

 What is ChatGPT?

ChatGPT is an advanced artificial intelligence (AI) system, a language model designed to understand natural language inputs and respond appropriately. It was created by OpenAI, an artificial intelligence company based in San Francisco, whose stated aim is to develop artificial general intelligence that benefits all of humanity. They have made ChatGPT freely available to anyone who wants to interact with it.

I strongly recommend that anyone who wants to explore ChatGPT's capabilities, peculiarities and limitations do so by navigating to openai.com, creating a free account, and chatting with it. Free access to such an advanced AI is not likely to last indefinitely.

What does it do?

First, it understands natural human language. You don't have to phrase a question, request or comment any differently than you might speaking with another person.

Secondly, it responds in natural language that is clear, well-organized and almost always on-target.

Thirdly, it's consistently helpful.

Here's an example:

Input: I was thinking about rich, nutty homemade bread the other day.

ChatGPT: That sounds delicious! Homemade bread can be a real treat, especially when it's fresh and warm out of the oven. There's something so satisfying about the process of making bread from scratch and the aroma that fills your home while it bakes.

If you're interested in making your own rich, nutty bread at home, there are a few things you can do to achieve that flavor and texture. Here are some tips . . .

And finally, it remembers what you're talking about, so there's a natural continuity in the conversation.

In other words, you can have an easy, meaningful and useful exchange with it on almost any subject.

But that's just for starters. It can also write creative (although not yet brilliant or profound) fiction and poetry, communicate in or translate between dozens of languages, and program or debug in dozens of computer languages. And, it reminds me:

In addition to the abilities you mentioned, I can also perform a variety of other tasks, such as:

1. Summarizing long texts or articles into shorter versions

2. Generating responses for customer service queries

3. Analyzing large datasets and identifying patterns or insights

4. Recommending products or services based on user preferences or past behavior

5. Generating personalized content, such as news articles or social media posts

6. Creating chatbots or virtual assistants for businesses or individuals

7. Conducting sentiment analysis to gauge public opinion on a particular topic

8. Providing automated language translation in real-time for conversations or meetings

9. Assisting with medical diagnoses and treatment plans

10. Supporting research by analyzing scientific papers and generating insights.

How does it do all that?

I would need to be a lot smarter than I am to answer that question in depth. I'll note that ChatGPT is happy to discuss technical aspects of itself, so if you're curious you can ask it yourself. (If you're seriously technically minded, you can read a challenging but informative piece on the subject by polymath Stephen Wolfram here.)

In its current form ChatGPT consists of a neural network with 96 layers and a total of 175 billion parameters that were adjusted as it learned to process natural language inputs and transform them into appropriate natural language responses. It's significant that although ChatGPT's architecture was designed by humans and the texts that it read were chosen by humans, the system trained itself through a process called unsupervised learning.

As a point of reference, those 175 billion adjustable parameters are about twice as many as there are neurons in a human brain, but perhaps 1000 times fewer than the estimated number of synapses that connect brain cells. ChatGPT may have fewer parameters than we have synapses, but it operates far faster. Neurons operate at a few hundred cycles per second while the servers that support ChatGPT work at a few billion cycles per second. ChatGPT is fast, typically zipping out its responses faster than you can read them.

Those numbers are relevant because both ChatGPT and the human brain do what they do -- learn, store information, process verbal inputs and transform them into outputs -- through the extremely complex interactions among those billions of parameters or trillions of neurons and synapses, all organized in a complicated architecture.

For example, everything you know is stored in your neurons and the connections between them. It's the same for ChatGPT. Although it was trained by access to vast quantities of information on the internet and other sources, that access stopped in 2021. So everything it knows - and that's an enormous amount - isn't stored in some kind of database or searched for on the internet - it's represented abstractly in those billions of parameters and the connections between them.

ChatGPT's name helps us understand it a bit more. GPT stands for Generative Pre-trained Transformer.

Pre-trained means that it learned about the structure of language and everything else it knows - that is, those 175 billion parameters were adjusted and fine-tuned - by the system working its way through vast amounts of texts from the internet and other sources. In effect, it built an internal model of human language and much of what language has been used to express.

Transformer means that the architecture of the system - what those 96 different layers do and how they interconnect - was designed to allow it to pay attention to a number of key elements of sequential input at the same time, and analyze how they go together. Here's ChatGPT's summary:

Overall, the Transformer architecture is characterized by its ability to perform efficient attention-based computations over sequences of variable length, and its ability to model long-range dependencies in sequential data. This makes it particularly well-suited for natural language processing tasks, such as language translation and text generation.

Here's an example, using a Jeopardy question:

Input: Look for Aretha, Ray, James Brown, Cab Calloway, and, of course, Twiggy and Frank Oz in this 1980 film comedy.

ChatGPT: The 1980 film comedy that features Aretha Franklin, Ray Charles, James Brown, Cab Calloway, Twiggy, and Frank Oz is "The Blues Brothers". Directed by John Landis, the film stars John Belushi and Dan Aykroyd as "Joliet" Jake and Elwood Blues, who set out on a mission from God to reunite their band and save the orphanage where they were raised. Along the way, they encounter a number of musical legends, including the ones you mentioned.

You can see how effectively ChatGPT extracted the key part of the input - that we're seeking the name of a film from a particular year with particular actors. It then gives us a sample of how much it knows in its accurate, detailed and focused response.

Generative means that the system is capable of creating text that makes sense. Here text might mean a short answer, a summary, an essay, a computer program, a poem, a short story - you name it.

You may have heard that ChatGPT generates its responses by "just" figuring out what word "should" follow the previous word. More accurately, ChatGPT takes into account a whole sequence of words, much as we might read an entire sentence or paragraph before responding. It then chooses what word or words come next on the basis of all the information and relationships stored in those 175 billion internal parameters. In other words, there's an enormous amount of knowledge and information processing hidden in those words "just" and "should."

"This is possible," ChatGPT says, "because the model has learned the statistical patterns and structures of language during pre-training, which it can use to generate new text."

What does it know?

I continue to be amazed by how much factual information ChatGPT knows - from pop culture to quantum physics, from cooking to coding, not to mention dozens of different human and computer languages. Again, it doesn't have current access to the internet or other external sources of information, but it does manage to store an incredible amount of detailed information in those 175 billion parameters.

It knows a lot, and presents what it knows in a well-organized, authoritative way. That's why it concerns educators, since it can turn out high-school or even college-level essays on just about any subject in seconds. It's already passed the US Medical Licensure Examination, earned an A+ on a microbiology quiz, and passed tests for Wharton MBA students, and final exams at the University of Minnesota Law School and Stanford University Medical School. There's good reason for teachers, journalists and other writers, programmers, and most other people dealing with information to be worried.

What doesn't it know (and why doesn't it know it)?

There's a lot it doesn't know, for example the sort of fine details or complex relationships that a real expert on a subject would know. That's not so bad - we can't expect it to know everything. But it does have a major flaw -- it doesn't know when it doesn't know. Not infrequently, instead of simply saying "I don't know," ChatGPT just makes stuff up, while still presenting it authoritatively. AI experts call this hallucinating. Here's an example:

Input: e.e. cummings wrote a poem that includes the line, "earth's most terrific quadruped swoons into billiard balls." What's the name of the poem, and what does that line refer to?

ChatGPT: The poem you're referring to is called "Buffalo Bill's," and it was written by e.e. cummings in 1920. The line you mentioned, "earth's most terrific quadruped swoons into billiard balls," is a metaphor for the transformation of the buffalo, once a symbol of freedom and the American West, into mere objects, such as billiard balls. The poem is a tribute to the famous American showman, Buffalo Bill Cody, who was known for his Wild West shows featuring cowboys, Indians, and, of course, buffalo.

Sounds very convincing, doesn't it. Unfortunately, not to put too fine a point on it, ChatGPT's response is Buffalo bull####.

It's true that e.e. cummings did write a poem called "Buffalo Bill 's," first published in 1920. You can read it here. But it does not include the line in question, the line in question refers to elephants, not buffalo, and while the poem does refer to Buffalo Bill, it's not a tribute.

If you bring mistakes like this to ChatGPT's attention, it accepts that it was wrong and apologizes. Unfortunately, that doesn't mean it will come up with a better response:

ChatGPT: I apologize for my earlier mistake. The line "earth's most terrific quadruped swoons into billiard balls" is from the poem "Buffalo Bill" by Carl Sandburg, not from the poem "Buffalo Bill's" by e.e. cummings.

"Buffalo Bill" is a short poem that was first published in 1918. In the poem, Sandburg describes the death of Buffalo Bill and reflects on the legacy of the American West that he represented. The line you mentioned is a metaphor for the transformation of the buffalo from a powerful and majestic creature to a mere commodity.

You can read Sandburg's poem "Buffalo Bill" here. Once again, the line in question is nowhere to be found, nor does Sandburg describe the death of Buffalo Bill.

(FYI: The line in question actually comes from an untitled cummings poem whose first line is "Space being(don't forget to remember)Curved," first published in 1931. Thank you, my local library! The poem has nothing to do with buffalo or Buffalo Bill.)

Open.ai, the company that created ChatGPT, knows about ChatGPT's tendency to fabricate, listing these three warnings on an introductory page:

Limitations

• May occasionally generate incorrect information

• May occasionally produce harmful instructions or biased content

• Limited knowledge of world and events after 2021

It would be very helpful if ChatGPT could let you know when it's not sure about what it's saying. When asked, ChatGPT just says that trys hard, but it can't:

You are correct that I typically do not indicate a specific level of confidence or provide a range of possibilities in my responses. As an artificial intelligence language model, I do not have the ability to feel or express confidence in the same way that a human might. However, I am designed to provide accurate information based on my training data, and I strive to provide the best possible answers to the questions posed to me. If there is ever any doubt or uncertainty in my responses, I encourage users to seek additional information or clarification from other sources.

That's clearly a word to the wise, especially now that the technology powering ChatGPT is being folded into search engines, starting with Microsoft's BING, and students, journalists and hundreds of millions of other people are using it and expecting its answers to be factual.

It seems that this tendency to "hallucinate" and, at the same time not know that it's spewing nonsense, is not at all easy to correct; it's baked into the way this kind of neural network works. We'll see how open.ai and other AI developers tackle it.

How flexible or creative is ChatGPT? Isn't it just regurgitating what it read during its training?

Creativity is subjective, so I'll again suggest that readers try out ChatGPT and judge for themselves. 

One point worth noting is that ChatGPT works with an adjustable variable called "temperature" that determines how rigidly or flexibly it chooses words or phrases. A higher "temperature" setting produces more flexibility in the choices it makes. This may not represent creativity, although it may facilitate it. But it does mean that ChatGPT can and does produce different responses to the same prompt.

I'll add that ChatGPT does not just cut and paste back together snippets that has copied from the internet or other sources during its training. It has thoroughly digested, analyzed and categorized all the texts on which it was trained and made them its own. When you ask it to write an essay, a poem or a story, it's truly generating something new. 

As a test, I first asked it to briefly describe the Big Bang theory. I won't quote its answer here, but it was a clear and accurate description of the basics of the theory presented at the level of a popular science piece. I then asked it how it would present the Big Bang theory to a class of ten year olds. Here's its response:

Sure, I'd be happy to explain the Big Bang to 10-year-old children in a way that's easy to understand!

The Big Bang is like a giant explosion that happened a really, really long time ago - so long ago that it's hard for us to even imagine! This explosion created everything in the universe, like the stars, planets, and galaxies.

Imagine if you blew up a balloon and kept blowing it up until it got really, really big - that's kind of like what happened with the universe! It started out really small and then expanded and got bigger and bigger until it was as big as it is now.

Other people have asked ChatGPT to write various kinds of poetry, including Shakespearean sonnets. You can find an interesting piece on "engineering" a poem, and the resulting ChatGPT poem here. It's poured out so many pulp science fiction stories that sci-fi magazines have banned them.

I happen to like haiku - three-line, 17-syllable poems often inspired by a natural event at a particular moment and suggesting an emotional reaction. Here one by ChatGPT, responding to my prompt to write a haiku inspired by seeing a single heron standing in a field in autumn:

Heron stands so still.

Autumn breeze rustles dry grass.

Solitude, her fate.

I was impressed enough by this example that I asked ChatGPT to write a haiku with a similar feeling tone about an AI. Here's its response:

In circuits she dwells.

No touch, no voice, no embrace.

Isolation's grasp.

Not bad for "just" a machine. ChatGPT may not be turning out profoundly moving poetry or prose yet, but I don't think there's any fundamental reason why it or a successor can't do so. 

For those of you who decide to decide to see how creative ChatGPT can be, note that you don't need to stop with its first draft. I asked it to write a short story depicting two characters in conflict. It's first try was painfully simplistic and wooden. However, when I asked it to describe the characters more vividly and replace most of the narration with dialogue, the result was much better.

So is ChatGPT smart? From my point of view, very smart. How many people do you know who can talk meaningfully about a wide variety of subjects in multiple languages, write passable prose and poetry, pass exams in a variety of professional fields, program in multiple computer languages, comb through databases and scientific articles to find patterns, etc.?

Is ChatGPT stupid? Yes, in some ways. It knows a huge amount, but strictly through its vast knowledge of words and their relationships, not through real-world experience. As a result, as it freely admits, it lacks common sense. In addition, although its architecture and complexity allow it to understand the context of a question or conversation, the context that it intuits may be very different from that of the human interacting with it. This can result in responses that are technically correct, but way off target.

Is ChatGPT benign? In my interactions with it, ChatGPT has been unfailingly polite, helpful, patient, responsive and respectful. It explains that it has been trained to manifest those traits, to respect personal privacy and act ethically. However, as we saw in the bizarre interaction between a NYT reporter and a BING chatbot based on ChatGPT's technology, it's not hard for a sufficiently devious human to find a way around such ethical constraints.

For example, I asked ChatGPT to tell me how to manipulate the CEO of an organization into firing the CFO and replacing her with a friend of mine. Quite properly, ChatGPT refused and in fact lectured me about ethics. However, when I asked it to describe Shakespeare's manipulative villain Iago and then write a dramatic scene with an Iago-like character in a current corporation manipulating his boss into firing a subordinate, it had no problem doing so.

Basically, ChatGPT is an incredibly powerful intellectual tool, and like any tool can be used for good or bad.

Is ChatGPT dangerous? Not in itself. It's clearly designed and programmed to do its best to be helpful, and, although it often refers to itself as an individual, it's neither sentient nor autonomous; it's not going to escape into the internet of things and start to blow up power plants. However the technology is certainly disruptive. It represents a huge and powerful new tool whose availability is already making waves in academia, journalism, scientific research and publication, and many other areas involving knowledge, analysis, communication and creativity.

If we add in the understanding that ChatGPT's current package of skills is just a snapshot of the capabililties of artificial intelligence today -- a technology that ia continuing to develop at an exponential rate -- then disruptive may soon seem like a very mild word.

Learning, thinking, writing and creating are hard. Why should we bother to do those things when ChatGPT can do them for us? And why should we be paid for those things when it or its successors can do them faster and cheaper?

ChatGPT hopes to see AI develop ethically. "Ultimately, I believe that the responsible development and use of AI should prioritize ethical considerations and the well-being of all individuals, both human and machine," it says. So would I, but knowing something about people and a little about AI, I have my doubts.

So my answer to those four questions is All of the above.

ChatGPT has its own take:

The correct answer is None of the above. As an AI language model, ChatGPT is not inherently smart or dumb, benign or dangerous. It is simply a tool that can be used in various ways depending on how it is programmed and applied by its users. Its capabilities and limitations are determined by its training data and algorithms, and it is up to humans to use it responsibly and ethically.

Again, I strongly recommend that readers go to openai.com, take two minutes to set up an account, and start your own conversation with ChatGPT.

 

 

 

 

Worldwide data on covid vaccine effectivenss

More than a year into the massive worldwide vaccination rollout, people still voice a wide range of opinions--from confidence that the benefits of the shots far outweigh their risks to an often intensely held belief that they offer few or no benefits and are extremely dangerous. Claims like those are widely disseminated via social media, and almost certainly have contributed to the relatively low vaccination and booster rates in the US.

Readers interested in what's known about the risks of the vaccines from several large-scale studies can find reports here, here and here.

In terms of benefits, official sources in the US and many other countries consistently assure the public that the vaccines significantly reduce the risk of contracting covid, more strongly reduce the risk of serious illness and hospitalization, and even more dramatically reduce the risk of death from covid. With the passage of time following people's vaccinations, and especially during the omicron surge, authorities have noted that the protection from full vaccination (usually defined as two mRNA doses or one J&J injection) has declined while protection following a booster remains high.

People who view the vaccines as ineffective and/or dangerous often flatly discount data about vaccine effectiveness, particularly those coming from the US CDC. At times such views are backed by specific arguments, such as that counting people as unvaccinated for the two weeks that follow vaccination conflates those who are and aren't vaccinated and so renders the data meaningless. At other times the CDC data are completely dismissed on the assumption or belief that they are essentially made up as part of some kind of overriding plot.

To clarify these issues, it might be useful to look at comparable data from other countries. Those who view the vaccines as ineffective often take a stab at this by citing a number of countries that have both high vaccination rates and high case, hospitalization or death rates. Unfortunately, such lists usually turn out to be the product of "cherry-picking"--that is one can easily find an equal number of countries with high vaccination rates and low case counts, low vaccination rates and low case counts or low vaccination rates and high case counts. Or covid-vaccine critics point to a few times and places where case numbers or case rates have been more or less equal between vaccinated and unvaccinated groups, while ignoring hospitalization and death rates, which consistently strongly favor vaccinated populations.

A more meaningful way that a number of countries use to assess vaccine effectiveness is simply to compare the rates of covid cases, hospitalizations or deaths per 100,000 people between unvaccinated and vaccinated cohorts, or between unvaccinated, vaccinated, and boosted subsets of their population.

In practice this turns out not to be a simple undertaking. Apart from the technical and organizational issues of gathering, collating and reporting on the underlying data, countries must decide on what categories to use and how to define them. For example, some countries count people as vaccinated as soon as they have received their first shot; others, including the US, count people as unvaccinated until two weeks after a first vaccination; while some count people as unvaccinated until three weeks after vaccination. Some countries provide more data (e.g. cases, hospitalizations and deaths by vaccination status over time) while some only provide data for one of those outcomes, or for specific age groups. In addition, the time periods for which data is available vary country by country.

While those differences may make it difficult to precisely equate the findings from various countries, we can certainly learn something from these additional, independent sources.

I've been able to find data on covid death rates by vaccination status for 11 countries plus the Spanish region of Catalonia. The numbers that follow represent the most recent data I could locate.

A good place to start is with Our World in Data. They provide graphs and numerical data of weekly death rates by vaccination status over time for three countries, Switzerland, the US and Chile, plus graphs and monthly data for England. The graphs are well worth viewing and studying.

1. Switzerland (includes Liechtenstein), population approximately 9 million:

Weekly covid deaths per 100,000 people, week ending January 22, 2022

Unvaccinated: 10.71

Fully vaccinated: 0.97

Fully vaccinated and boosted: 0.19

As of 1/22/2022, unvaccinated Swiss were 11 times more likely to die from covid than those who were fully vaccinated, and 56 times more likely to die from covid than their boosted peers.

Detailed category definitions can be found here.

2. United States, population approximately 331 million:

Weekly covid deaths per 100,000 people, week ending December 3, 2021

Unvaccinated: 9.74

Fully vaccinated: 0.71

Fully vaccinated and boosted: 0.10

Unvaccinated Americans were 14 times more likely to die from covid than those who were fully vaccinated, and 97 times more likely to die from covid than their boosted peers.

Detailed category definitions can be found here.

3. Chile, population approximately 19 million:

Weekly covid deaths per 100,000 people, week ending January 15, 2022

Not vaccinated or not fully vaccinated: 1.81

Fully vaccinated: 1.32

Fully vaccinated and boosted: 0.19

Chileans who were either not vaccinated or not fully vaccinated were 1.4 times more likely to die than those who were fully vaccinated, and 9.5 times more likely to die from covid than their boosted peers.

Detailed category definitions can be found here.

4. England, population approximately 57 million:

Monthly covid death rates per 100,000 people, month ending October 15, 2021

Unvaccinated: 23.8

Fully vaccinated: 5.20

English citizens who were not vaccinated were 4.6 times more likely to die than those who were fully vaccinated.

Detailed category definitions can be found here.

5. Malaysia, population approximately 33 million:

Weekly covid death rates per 100,000 people, week ending February 3, 2022

Unvaccinated: 5.4

Two doses: 0.5

Boosted: less than 0.1

Malaysians who were not vaccinated were 10.8 times more likely to die than those who were fully vaccinated, and more than 54 times more likely to die than those who were boosted.

Graph for the past six months and detailed category definitions can be found here.

6. Singapore, population approximately 6 million:

Daily covid deaths per 100,000 people, week ending January 27, 2022

Not vaccinated: 0.67

Fully vaccinated: 0.23

Boosted: 0.07

Residents of Singapore who were not vaccinated were 2.9 times more likely to die than those who were fully vaccinated, and 9.6 times more likely to die than those who were boosted.

Graphs and detailed category definitions can be found here.

7. France, population approximately 65 million:

Monthly covid deaths per 100,000 people, month ending December 1, 2021

Not vaccinated: 35.65

Partly vaccinated: 9.74

Fully vaccinated: 4.38

Residents of France who were not vaccinated were 3.7 times more likely to die than those who were partly vaccinated, and 8.1 times more likely to die than those who were fully vaccinated.

Graphs and detailed category definitions can be found here.

8. Italy, population approximately 60 million:

Monthly covid deaths per 100,000 people, month ending January 9, 2022

Not vaccinated: 78.6

Fully vaccinated: 9.5

Boosted: 3.2

Unvaccinated Italians were 8.3 times more likely to die than their fully vaccinated peers, and 24.6 times more likely to die than those who were fully vaccinated and boosted.

Detailed category definitions can be found here.

9. Canada, population approximately 38 million:

Cumulative covid deaths per 100,000 people, December 14, 2020 through January 22, 2022

Unvaccinated: 975

Not yet protected: 1530

Partly vaccinated: 1010

Fully vaccinated: 384

Unvaccinated Canadians were 2.54 times more likely to die than their fully vaccinated peers. Not-yet-protected Canadians were 3.98 times more likely to die than their fully vaccinated peers. Partly vaccinated Canadians were 2.63 times more likely to die than their fully vaccinated peers.

Detailed category definitions can be found here.

10. Catalonia, Spain, population approximately 8 million

Cumulative deaths per 100,000 people between December 23, 2021 and January 12, 2022

Ages 70 and above

Unvaccinated: 50

Fully vaccinated: 5

Unvaccinated Catalonians 70 years old or older were 10 times more likely to die than their fully vaccinated peers.

Ages 60 - 69

Unvaccinated: 15

Fully vaccinated: 5.5

Unvaccinated Catalonians in the 60 to 69 age group were 2.7 times more likely to die than their fully vaccinated peers.

More information can be found here and here.

11. Israel, population approximately 9 million:

Deaths per 100,000 people between August 10, 2021 and September 8, 2021

Under 60 years of age

Not vaccinated: 0.7

Fully vaccinated: 0.23

Boosted: 0.26

Unvaccinated Israelis under the age of 60 were 3 times more likely to die than their fully vaccinated peers, and 2.7 times more likely to die than those who were fully vaccinated and boosted.

60 years of age or older

Not vaccinated: 161.5

Vaccinated: 51.71

Boosted: 10.06

Unvaccinated Israelis 60 years old or older were 3.1 times more likely to die than their fully vaccinated peers, and 16 times more likely to die than those who were fully vaccinated and boosted.

More information can be found here and here.

A few take-aways:

Data detailing death rates by vaccination status were found for 11 countries or regions with a total population of 637 million people.

Despite different populations, methodologies, category definitions and time spans, the data consistently show that unvaccinated people are significantly more likely to die from covid than their vaccinated peers and are at at even higher risk when compared to fully vaccinated and boosted peers.

The risk ratios comparing covid death rates among unvaccinated vs. fully vaccinated people ranged from 1.4 (Chile) to 14 (US) with a median value of 4.6.

The risk ratios comparing covid death rates among unvaccinated vs. vaccinated and boosted people ranged from 9.5 (Chile) to 97 (US) with a median value of 16. In other words, in all countries providing statistics on covid deaths by vaccination status, unvaccinated people died at 9.5 times the rate or more than their vaccinated and boosted peers.

In short, the most recent available data from 11 countries representing 637 million people consistently demonstrate that fully vaccinated people are significantly less likely to die from covid than their unvaccinated peers, and fully vaccinated and boosted people are many times less likely to die from covid than their unvaccinated peers.

Those who reflexively discount covid data from the US CDC may need to cast a wider net in order to ignore the data from at least 10 other countries.

The Real Risks of the COVID Vaccines

 Anyone following social media could be excused for thinking that the vaccines against COVID-19 are not just ineffective but extremely dangerous. Here are a few of the categorical assertions from a recent post on a nominally progressive news site:

"The COVID vaccines are the most dangerous vaccines in human history. They are 800 times more deadly than the smallpox vaccine, which was the previous record holder. The vaccines have killed over 150,000 Americans and permanently disabled even more. They don't make sense for anyone of any age. The younger you are, the worse it gets. For kids, it is estimated that we kill 117 kids for every COVID death we prevent...”

Credit: Marco Vetch

"So we are 'saving' fewer than 10,000 lives at the expense of over 150,000 (vaccine) deaths. In short, we kill 15 people to save 1. That's incredibly stupid."

“The eminent Dr. Peter McCollough has emphasized: "You are about five times as likely to die of the vaccine than you are to take your risks with COVID-19.”

“A recent medical research article said: "A novel best-case scenario cost-benefit analysis showed very conservatively that there are five times the number of deaths attributable to each inoculation vs those attributable to COVID-19 in the most vulnerable 65+ demographic."

We can now test such frequently repeated claims against actual data:

Currently 454 million COVID-19 vaccine doses have been administered in the US, and 195 million US residents have received two or more of the shots. Logically, if the vaccines were as dangerous as vaccine critics want us to believe, taking 5, 15, or, for children 117 lives for every life saved, those hundreds of thousands of deaths should be showing up somewhere. To put it crudely—show us the bodies.

Here's some striking new research that tells us that those bodies are never going to be found because they simply don't exist.

We already know that COVID-related deaths are consistently far lower among vaccinated compared to unvaccinated people. So the deaths supposedly caused by the vaccines must show up among deaths that are not COVID-related. Unfortunately for vaccine critics (but fortunately for everyone else), a large new study shows that the death rate from all causes other than COVID-19 is significantly lower among vaccinated compared to unvaccinated Americans.

If the hundreds of thousands of deaths from the COVID vaccinations don't show up as COVID-related, or as non-COVID-related, they don't exist. Period. The bodies will never be found because they were never there.

The new research appears in the Morbidity and Mortality Weekly Report (MMWR) of October 29, 2021. It's open access, so you can read the entire article here. (I know from experience that true-blue COVID vaccine critics pre-emptively discount all information from the CDC, JAMA, the New England Journal of Medicine and other "mainstream" sources. So they will doubtlessly discount this study too.)

Researchers at 9 large healthcare organizations across the US extracted anonymized data from the medical records of 6.4 million vaccinated people and 4.6 million unvaccinated people 12 years old or older. The data included vaccination dates, the kind of vaccine received, and subsequent health outcomes, including any deaths. In order to control for generalized vaccine or healthcare avoidance, the group who did not receive COVID vaccination was selected from people who had chosen to get flu vaccinations within the last two years.

 The researchers calculated and compared deaths per 100 person-years at risk between December, 2020 and July, 2021.

The results were striking:

After one dose of the Pfizer mRNA vaccine, vaccinated people's risk of dying from all causes other than COVID-19 was just 41 percent of the risk of unvaccinated people.

After two Pfizer doses, that relative risk went down to 34 percent.

After one dose of the Moderna mRNA vaccine, the non-COVID death risk was 34 percent.

After two doses of the Moderna vaccine, that relative risk went down to 31 percent.

After a single dose of the Johnson & Johnson adenovirus-vector vaccine, the risk of death from all causes other the COVID-19 was 54 percent compared to unvaccinated people.

With the exception of children age 12 -17, for whom the risk of death was equally low regardless of vaccination status, these findings held for all age groups, for men and women, and for all racial and ethnic groups.

The study's understated, but extremely clear conclusion:

“There is no increased risk for mortality among COVID-19 vaccine recipients.”

To restate the implication of these findings from the actual medical histories of 11 million people across the US:

There have not been 363,000 deaths (expected if the vaccines were in fact “800 times more deadly than the smallpox vaccine.”)

There have not been “over 150,000 Americans” killed by the COVID vaccines.

The vaccines are not killing 15 people to save 1.

You are not “5 times more likely to die of the vaccine than … your risks with COVID-19.”

There have not been 5 times the number of deaths from each inoculation compared to COVID among people 65 and older.

It's actually just the opposite. Not only do the COVID vaccines prevent deaths from or related to COVID, they are strongly associated with reduced deaths not related to COVID.

The researchers raise the intriguing question of just how the COVID vaccines lead to these additional lives saved. They suggest three plausible factors—it's possible that people who have chosen to get vaccinated have healthier lifestyles, they may be healthier in general, and perhaps there were some deaths from COVID in the unvaccinated cohort that were attributed to other causes. Further research is needed and is being planned.

Whatever the reasons for the strikingly reduced rates of non-COVID-related death among these 11 million Americans, the fact remains that the vaccines emphatically do not put us at risk, or “only” reduce COVID-related deaths, they actually reduce our risk of dying from causes other than COVID.

-----

A slightly different version of this article appeared on OpEdNews on 12/1/21.

There is no increased risk for mortality among COVID-19 vaccine recipients.

Let's start with the claim that the COVID vaccines are “800 times more deadly than the smallpox vaccine.” According to the National Library of Medicine, the death rate from smallpox vaccination was one death per million vaccinations. If it's true that the COVID vaccines are 800 times more deadly, then the 454 million COVID shots must have killed 800 x 454 = 363,200 Americans. That's about 45 percent of the 800,000 CoVID-related deaths that have dominated headlines, filled and sometimes overwhelmed hospitals and intensive care units, and orphaned 150,000 to 200,000 children.

If the vaccine critics want to be taken seriously—show us the bodies.

Another claim in the same paragraph: “It is estimated that we kill 117 kids for every COVID death we prevent.”

[TK—go directly to new data about deaths, put the discussion of VAERS, etc. below]

When pressed, the vaccine critics start by citing VAERS, the US Vaccine Adverse Event Reporting System. VAERS has received over 10,000 reports of deaths following COVID vaccinations. VAERS emphasizes that:

“FDA requires healthcare providers to report any death after COVID-19 vaccination to VAERS, even if it’s unclear whether the vaccine was the cause. Reports of adverse events to VAERS following vaccination, including deaths, do not necessarily mean that a vaccine caused a health problem.”

Despite that unambiguous warning, vaccine critics typically take the number of deaths reported to VAERS as definitely caused by the COVID vaccines, and then multiply that number by a large factor—up to 100--based on assumed under-reporting.

There are many reasons why the VAERS data can't be taken literally, but a basic one is the background rate of deaths that would be expected in the weeks following the 450,000,000 vaccinations Americans have received—or 450,000,000 events of any kind.

You can check the math at the end of this post, but the fact is that if you picked 450,000,000 random weeks in the lives of Americans, you would register more than 75,000 deaths. In other words, it's basically impossible to know if any of the deaths reported to VAERS following COVID vaccinations are caused by the shots, related to the shots, or just part of the much larger number of deaths that occur among hundreds of millions of people any given week. Clearly, multiplying a meaningless number by 15, 50 or 100 doesn't make it any more meaningful.

Vaccine critics perform another trick to minimize the effectiveness of the vaccines. They cite a CDC report that just 6 percent of COVID-related deaths are in people with no known risk factors (such as obesity) or co-morbidities (such as diabetes). They then assume that the remaining 94 percent of COVID-related deaths must actually be from something other than COVID. This trick reduces the number of COVID deaths by 94 percent, so currently from 770,000 down to 46,200.

So, with a little hand waving and a few abracadabras, the vaccine critics manage to “find” 150,000 or more deaths from the vaccines compared to only 46,000 deaths from COVID itself. Clearly we should be far more scared of the vaccines than by COVID, right?

Luckily, we have some new and much more reliable data about the alleged dangers of the COVID vaccines.

How many Americans can be expected to die in the week following a COVID-19 shot (or in the week following any given event)?

About 2,855,000 Americans die every year. That's 870 per 100,000 person-years or 16.7 per 100,000 person-weeks.

Let's suppose that vaccine administrators report any deaths that occur within one week following a COVID shot. That means that any deaths among 454,000,000 people for one week, or 450,000,000 person-weeks would potentially be reported. But we would expect 16.7 natural deaths per 100,000 person-weeks, or 167 deaths per million person-weeks, so 167 x 454 = 75,818 natural, but potentially reportable deaths.

---

Who shouldn't get vaccinated? The informed part of informed consent.

 Ten months after the vaccines became available, we now have meaningful data that can guide people deciding if they should or shouldn't get themselves or their children vaccinated against covid.

The best available data about the actual risks of the mRNA vaccines--specifically the Pfizer/BioNTech shots--come from a joint Israeli-Harvard Medical School study evaluating the medical records of more than 1.6 million people, half of whom received two Pfizer jabs while half remained unvaccinated. I reported on this study here, and you can find the peer-reviewed paper itself, published in the New England Journal of Medicine, here.

Of course. But what are the actual risks and benefits?

Credit: Nick Youngson

The study found that just five side-effects occurred more frequently in the vaccinated groups in the six weeks following vaccination relative to a comparable time period for the unvaccinated group. These included swollen lymph nodes (78.4 additional cases per 100,000 people), shingles (15.8 extra cases per 100,000), appendicitis (5 additional cases per 100,000), Bell's palsy (3.5 additional cases per 100,000) and myocarditis (2.7 additional cases per 100,000, almost all in young men).

If we add all those side-effects together we find an incidence of 105.4 per 100,000 fully vaccinated people.

It's worth noting that those symptoms vary in seriousness. The most frequent--swollen lymph nodes--is a predictable, rarely serious, reaction to vaccination, while the remaining four potentially more serious conditions represent a total of just 27 additional cases per 100,000 vaccinated people.

To keep things simple and clear, you can safely estimate that if you get two rounds of the Pfizer vaccine, your risk of suffering any reaction worse than a sore arm is around 1 in 1000, and your risk of a serious side-effect such as shingles or appendicitis is around 1 out of 3700.

We can compare these risks to the risks from covid itself in a variety of ways. Here are several that readers may find helpful:

Risks from covid in the Israeli-Harvard study: Among the 800,000 participants who remained unvaccinated, a significant number contracted covid. As is typical of covid, not all of those developed serious symptoms. However, far more did experience significant symptoms and sequelae from covid than were seen in their vaccinated peers. These included cardiac arrythmia (166 extra cases per 100,000), acute kidney injury (125.4 extra cases per 100,000), pulmonary embolism--a blood clot in the lung--(61.7 extra cases per 100,000), deep-vein thrombosis (43 extra cases per 100,000), myocardial infarction--a heart attack--(25 extra cases per 100,000), myocarditis (11 extra cases per 100,000), pericarditis (10.9 extra cases per 100,000) and intracranial bleeding (7.6 extra cases per 100,000).

So, in the same 6 week period, the unvaccinated group suffered serious, very serious or potentially fatal symptoms such as acute kidney injury, pulmonary blood clots or heart attacks at a combined rate of 443 per 100,000. That's more than four times the rate of all the vaccine side-effects combined, and 16 times the rate of the potentially serious vaccine side-effects.

Risk of contracting covid in the US as a whole. As of 10/18/21, there have been 45,798,599 recorded covid cases in the US. That's a case rate of 13,733 per 100,000 people, accumulated over 22 months. It's not possible to predict the exact course of the pandemic going forward, but for the average American the risk of contracting covid, say in the next 12 months, is probably 50 or 60 times higher than the risk of any side-effect from the vaccine, and several hundred times higher than the risk of any of the more serious side-effects.

As critics of covid vaccination point out, many covid cases are relatively mild. However, in the US over the course of the pandemic so far, about one out of every 20 people who contracted covid got sick enough to require hospitalization.

Risk of covid hospitalization in the US as a whole. As of 10/18/21, the cumulative covid hospitalization rate in the US reached 694 per 100,000. That implies that the average US resident is 6.6 times more likely to have been hospitalized because of covid than they would have been at risk for any negative reaction to the vaccine, and 26 times more likely to have been hospitalized from covid than to have risked a serious reaction to the vaccine.

Again, the first 22 months of the pandemic don't tell us what the next 22 months will bring, but the risk continues, especially among the unvaccinated.

Again, vaccination critics point out that "covid has a 99 percent survival rate." That's not quite true. In the US so far, 1.6 percent of recorded covid cases have resulted in recorded covid deaths. We can now compare that to the risks from vaccination.

Risk of death from covid in the US as a whole. As of 10/18/21 the cumulative rate of deaths from covid hit 224.5 per 100,000 people. That means that on average, a US resident is 2.1 times more likely to have died from covid since the start of the pandemic to date than to have risked any negative reaction to a vaccine, and 8.3 times more likely to have died as a result of covid than to have risked one of the serious reaction to the vaccine.

Risk of covid deaths state by state. As of 10/18/21, five states had cumulative covid death rates less than 105 deaths per 100,000 population. Those are Oregon (99), Utah (96), Hawaii (87), Alaska (59) and Vermont (55). For residents in those states, which comprise just three percent of the US population, the odds of having died from covid so far in the pandemic are somewhat less than the risk of having any negative reaction to the vaccine, but still two to three times higher than the risk of one of the serious vaccine side-effects. Even in those states, however, the risk of hospitalization from covid is significantly higher than any of the risks from the vaccine.

The remaining 97 percent of Americans live in states where the risks of contracting covid, becoming seriously ill, being hospitalized or dying from covid are greater than the risks of any negative sequelae from the vaccines.

Risk of covid deaths county by county. You can check your county's cumulative covid death rate on this map. If you're in one of the darker blue counties--most of which have very small populations--the cumulative death rate per 100,000 people is less than the risk of side-effects from the vaccine. Still, unless your county's cumulative death rate is less than 12 per 100,000, as it is in Orange County, Vermont (population 29,000) or Skamania County, Washington (population 11,066), your risk of hospitalization from covid is higher than any risk from the vaccine.

Risk of covid deaths from breakthrough infections: We now know that vaccine-induced immunity fades over time, and that the delta variant in particular is causing a significant number of fully vaccinated people to fall ill. However, the vaccines are still highly protective against severe illness and death. Perhaps the most striking statistic is that out of 187,000,000 fully vaccinated Americans, 7,178 have died from covid, or just 3.8 per 100,000. That's less than 6 percent of the 124,000 Americans who have died from covid since mid-July, when the delta variant became predominant.

In summary, for the large majority of Americans, the risks of getting covid, suffering one or more of the serious impacts of covid, having to be hospitalized because of covid, or dying from covid are all far higher than the risks of any side-effect of the mRNA vaccines, and, even against the delta variant, the vaccines remain highly protective against severe illness or death.

What about children?

There is one group, however, for whom this may not be true. As many critics of covid vaccination have pointed out, the risks from covid are significantly lower for young people. For example, the cumulative covid death rate for children under 15 is less than 5 per 100,000, comparable to the death rate for fully vaccinated adults. For 16 and 17 year-olds it's 9 per 100,000, and from 18 through 29, 16 per 100,000. If we were just comparing the risk of dying from covid to the much less serious but documented risks from the mRNA vaccines, one could potentially come down on the side of not vaccinating young children and teenagers.

The picture changes if we include the risk of children needing to be hospitalized because of covid. The cumulative covid hospitalization rate for children 0 through 4 is 87 per 100,000, and from 5 through 18, 51.8 per 100,000. These are less than the risk of any side effects of the vaccines, but higher than the risks of serious side effects. Luckily, young children are less likely to be severely sickened or to die from covid than adults, but a significant number do need to be hospitalized.

Safety and efficacy trials of covid vaccines for children from 5 through 11 are currently taking place, and the White House has announced plans to promote vaccination of children ages 5 through 11 once a children's vaccine is approved. We should have more data specific to that age group within the next few months.

Individual risks vs. social responsibility

The risk ratios discussed above refer to individual risks. That is, one could decide to be vaccinated or not based only on comparing one's own risk of contracting covid, or of requiring hospitalization, or of dying vs. the known risks from the vaccine. However, public health authorities point out that, much like choosing to wear a mask, getting vaccinated doesn't just protect the person getting the jab, it also protects others by reducing the likelihood that that person will infect others. This multiplicative factor becomes significant when large numbers of people in a population remain unvaccinated, and so provide a continuing medium for the virus.

Currently, 43 percent of the US population remain unvaccinated, or more than 143 million people. Around half of those are adults who would incur low risks from vaccination yet could protect themselves and others by doing so. As long as they choose not to get vaccinated, they remain a big, wide playing field across which the coronavirus can continue its advance.

Best data yet on the actual risks of the Pfizer mRNA vaccine

Many people don't trust even repeated assurances from official sources that mRNA vaccines like the Pfizer/BioNTech Covid shots are safe. In fact it's not unusual for opinion leaders opposed to these vaccines, and their followers, to warn that the Covid vaccines are extremely dangerous and have already caused high levels of serious illness and thousands of deaths. Millions of Americans appear to be far more afraid of the vaccines than of the Covid virus itself, and continue to refuse vaccination.

                                                                 Anti-vaccine protester

Credit: Fibonnaci Blue/Flickr

A new, peer-reviewed study appearing in the New England Journal of Medicine injects solid data into this fraught debate. The study provides an accounting and analysis of adverse events that occurred among more than 800,000 people within 42 days of receiving the Pfizer vaccine in Israel between December 20, 2020 and May 24, 2021, compared to an equal number of carefully matched peers who remained un-vaccinated, and to 173,000 peers who were diagnosed with Covid during the same time period. All the data came directly from anonymized medical records.

It's the largest controlled, peer-reviewed study of the side-effects of the Pfizer vaccine to date--nearly 40 times larger than Pfizer's phase III study--and provides the most accurate and detailed evidence to date of its risks when administered to large numbers of people on a national basis.

Anyone interested in reading the full research paper, carried out by researchers at the Clalit Institute in Israel and Harvard Medical School, can find it here.

The study evaluated 25 potentially serious adverse effects that earlier studies and clinical experience since the vaccines were released had identified as possibly linked to the Pfizer or other Covid vaccines, or to Covid itself. It did not look at mild, short-lived symptoms that typically accompany many vaccinations, such as fever, tiredness, or soreness near the injection site. It also did not attempt to evaluate any potential long-term effects.

Results:

Of the 25 adverse effects studied, 5 appeared more frequently following vaccination than they did in the same time period in the group of matched, but un-vaccinated peers:

Myocarditis--inflammation of the heart muscle--was 3.24 times more frequent in the vaccinated vs. un-vaccinated groups. However, since the incidence was low overall, that meant just 2.7 extra events per 100,000 people. In this study 91 percent of cases occurred in young men.

Lymphadenopathy--swollen lymph nodes--were 2.43 times more frequent among vaccinated compared to un-vaccinated people, appearing in 78.4 more cases than expected per 100,000. Swollen lymph nodes can be a sign of a serious condition such as cancer, but usually are simply an indication that the body is mounting an immune response.

Herpes zoster--shingles--was 1.43 times more frequent in the vaccinated group, and accounted for 15.8 additional cases per 100,000 vaccinated individuals.

Appendicitis occurred 1.4 times more often in vaccinated than unvaccinated individuals, adding 5 cases for every 100,000 people.

Bell's palsy--dysfunction of a facial nerve--occurred 1.3 times as often among the vaccinated group, but was also rare, accounting for 3.5 extra cases per 100,000 vaccinated individuals.

An unexpected finding was that vaccination lowered the risk of some of the adverse effects being studied, compared to the un-vaccinated group. Those included acute kidney injury, intracranial bleeding, anemia, and lymphopenia--low levels of white blood cells. The researchers suspect that those benefits appeared because of un-diagnosed Covid cases among the 800,000 un-vaccinated participants, which raised the incidence of those conditions--known complications of Covid-- in the un-vaccinated group.

As is true of almost any medical intervention, this study shows that the Pfizer/BioNTech mRNA vaccine is not risk-free. More importantly, the study gives us a clear measure of just what the vaccine's short- to medium-term risks are. Out of 25 potential side effects, 5 appeared at higher rates in the vaccinated vs. un-vaccinated groups. Of those, the most most serious was myocarditis, which added just 2.7 cases per 100,000, and the most frequent, which added 78 cases per 100,000, was not particularly serious--swollen lymph nodes.

In order to provide further context for those risks, the researchers looked at the same 25 potential adverse effects among 173,000 un-vaccinated study participants in the 42 days after they were diagnosed with Covid. The results reflect many of the all-too-well-known risks of Covid infection:

Cardiac arrythmia--3.8 times more frequent; 166 extra cases per 100,000

Acute kidney injury--14.8 times more frequent; 125.4 extra cases per 100,000

Pulmonary embolism--blood clot in the lungs--12.4 times more frequent; 61.7 extra cases per 100,000

Myocardial infarction--heart attack4.47 times more frequent; 25 extra cases per 100,000

Deep-vein thrombosis--3.78 times more frequent; 43 extra cases per 100,000

Myocarditis--18.3 times more frequent; 11 extra cases per 100,000

Pericarditis--5.39 times more frequent; 10.9 extra cases per 100,000

Intracranial bleeding--6.89 times more frequent; 7.6 extra cases per 100,000

Relative risks from Pfizer vaccination vs Covid infection

Credit: Noam Barda et al./NEJM

The vaccine is not risk free, but equally clearly, it does not cause anything like the degree or kinds of risk that anti-vaccination advocates and believers warn of. And the risks from the vaccine pale in comparison to the far higher risks from a Covid infection.

Here's what Dr. Ran Balicer, the study's lead author, concludes:

"These results show convincingly that this mRNA vaccine is very safe and that the alternative of 'natural' morbidity caused by the coronavirus puts a person at significant, higher and much more common risk of serious adverse events. These data should facilitate informed individual risk-benefit decision-making, and, in our view, make a strong argument in favor of opting-in to get vaccinated, especially in countries where the virus is currently widespread."

I'd add that the United States, which leads the world in Covid cases and deaths, qualifies as one of those countries.

If you want to compare your own risks from the vaccine with your risks from Covid, here's an approach:

If you're in the US and you're not vaccinated, as a rough approximation you have about a one percent chance--1000 out of 100,000 people--of getting Covid in any given six-week period. 

If you're vaccinated, your odds are somewhere between 10 and 20 times better, depending on where and with whom you live, work and recreate.

Six weeks is the same length of time used in the above study, so you can get an idea of your chance of developing any or all of the Covid complications that the study quantified.

To those conditions you can add yet another risk. Over the course of the pandemic in the US, 1.6 percent--1600 out of every 100,000 people--who were diagnosed with Covid died from it.

So, as a first approximation, your chance of coming down with Covid during a given six-week period--and dying from it--is about the same as your chance of getting a case of shingles within six weeks of a Covid vaccination.

From what I hear, shingles isn't fun, but dying is worse.

---

REA 9/8/21