Google

Everyone includes your mother and Los Angeles

PNRJ cognitive science, computer science, core principles, current events, ethics, futurism, heuristics and biases, politics, public policy , , , , , ,

Apr 28 JDN 2460430

What are the chances that artificial intelligence will destroy human civilization?

A bunch of experts were surveyed on that question and similar questions, and half of respondents gave a probability of 5% or more; some gave probabilities as high as 99%.

This is incredibly bizarre.

Most AI experts are people who work in AI. They are actively participating in developing this technology. And yet more than half of them think that the technology they are working on right now has a more than 5% chance of destroying human civilization!?

It feels to me like they honestly don’t understand what they’re saying. They can’t really grasp at an intuitive level just what a 5% or 10% chance of global annihilation means—let alone a 99% chance.

If something has a 5% chance of killing everyone, we should consider that at least as bad asthan something that is guaranteed to kill 5% of people.

Probably worse, in fact, because you can recover from losing 5% of the population (we have, several times throughout history). But you cannot recover from losing everyone. So really, it’s like losing 5% of all future people who will ever live—which could be a very large number indeed.

But let’s be a little conservative here, and just count people who already, currently exist, and use 5% of that number.

5% of 8 billion people is 400 million people.

So anyone who is working on AI and also says that AI has a 5% chance of causing human extinction is basically saying: “In expectation, I’m supporting 20 Holocausts.”

If you really think the odds are that high, why aren’t you demanding that any work on AI be tried as a crime against humanity? Why aren’t you out there throwing Molotov cocktails at data centers?

(To be fair, Eliezer Yudkowsky is actually calling for a global ban on AI that would be enforced by military action. That’s the kind of thing you should be doing if indeed you believe the odds are that high. But most AI doomsayers don’t call for such drastic measures, and many of them even continue working in AI as if nothing is wrong.)

I think this must be scope neglect—or something even worse.

If you thought a drug had a 99% chance of killing your mother, you would never let her take the drug, and you would probably sue the company for making it.

If you thought a technology had a 99% chance of destroying Los Angeles, you would never even consider working on that technology, and you would want that technology immediately and permanently banned.

So I would like to remind anyone who says they believe the danger is this great and yet continues working in the industry:

Everyone includes your mother and Los Angeles.

If AI destroys human civilization, that means AI destroys Los Angeles. However shocked and horrified you would be if a nuclear weapon were detonated in the middle of Hollywood, you should be at least that shocked and horrified by anyone working on advancing AI, if indeed you truly believe that there is at least a 5% chance of AI destroying human civilization.

But people just don’t seem to think this way. Their minds seem to take on a totally different attitude toward “everyone” than they would take toward any particular person or even any particular city. The notion of total human annihilation is just so remote, so abstract, they can’t even be afraid of it the way they are afraid of losing their loved ones.

This despite the fact that everyone includes all your loved ones.

If a drug had a 5% chance of killing your mother, you might let her take it—but only if that drug was the best way to treat some very serious disease. Chemotherapy can be about that risky—but you don’t go on chemo unless you have cancer.

If a technology had a 5% chance of destroying Los Angeles, I’m honestly having trouble thinking of scenarios in which we would be willing to take that risk. But the closest I can come to it is the Manhattan Project. If you’re currently fighting a global war against fascist imperialists, and they are also working on making an atomic bomb, then being the first to make an atomic bomb may in fact be the best option, even if you know that it carries a serious risk of utter catastrophe.

In any case, I think one thing is clear: You don’t take that kind of serious risk unless there is some very large benefit. You don’t take chemotherapy on a whim. You don’t invent atomic bombs just out of curiosity.

Where’s the huge benefit of AI that would justify taking such a huge risk?

Some forms of automation are clearly beneficial, but so far AI per se seems to have largely made our society worse. ChatGPT lies to us. Robocalls inundate us. Deepfakes endanger journalism. What’s the upside here? It makes a ton of money for tech companies, I guess?

Now, fortunately, I think 5% is too high an estimate.

(Scientific American agrees.)

My own estimate is that, over the next two centuries, there is about a 1% chance that AI destroys human civilization, and only a 0.1% chance that it results in human extinction.

This is still really high.

People seem to have trouble with that too.

“Oh, there’s a 99.9% chance we won’t all die; everything is fine, then?” No. There are plenty of other scenarios that would also be very bad, and a total extinction scenario is so terrible that even a 0.1% chance is not something we can simply ignore.

0.1% of people is still 8 million people.

I find myself in a very odd position: On the one hand, I think the probabilities that doomsayers are giving are far too high. On the other hand, I think the actions that are being taken—even by those same doomsayers—are far too small.

Most of them don’t seem to consider a 5% chance to be worthy of drastic action, while I consider a 0.1% chance to be well worthy of it. I would support a complete ban on all AI research immediately, just from that 0.1%.

The only research we should be doing that is in any way related to AI should involve how to make AI safer—absolutely no one should be trying to make it more powerful or apply it to make money. (Yet in reality, almost the opposite is the case.)

Because 8 million people is still a lot of people.

Is it fair to treat a 0.1% chance of killing everyone as equivalent to killing 0.1% of people?

Well, first of all, we have to consider the uncertainty. The difference between a 0.05% chance and a 0.015% chance is millions of people, but there’s probably no way we can actually measure it that precisely.

But it seems to me that something expected to kill between 4 million and 12 million people would still generally be considered very bad.

More importantly, there’s also a chance that AI will save people, or have similarly large benefits. We need to factor that in as well. Something that will kill 4-12 million people but also save 15-30 million people is probably still worth doing (but we should also be trying to find ways to minimize the harm and maximize the benefit).

The biggest problem is that we are deeply uncertain about both the upsides and the downsides. There are a vast number of possible outcomes from inventing AI. Many of those outcomes are relatively mundane; some are moderately good, others are moderately bad. But the moral question seems to be dominated by the big outcomes: With some small but non-negligible probability, AI could lead to either a utopian future or an utter disaster.

The way we are leaping directly into applying AI without even being anywhere close to understanding AI seems to me especially likely to lean toward disaster. No other technology has ever become so immediately widespread while also being so poorly understood.

So far, I’ve yet to see any convincing arguments that the benefits of AI are anywhere near large enough to justify this kind of existential risk. In the near term, AI really only promises economic disruption that will largely be harmful. Maybe one day AI could lead us into a glorious utopia of automated luxury communism, but we really have no way of knowing that will happen—and it seems pretty clear that Google is not going to do that.

Artificial intelligence technology is moving too fast. Even if it doesn’t become powerful enough to threaten our survival for another 50 years (which I suspect it won’t), if we continue on our current path of “make money now, ask questions never”, it’s still not clear that we would actually understand it well enough to protect ourselves by then—and in the meantime it is already causing us significant harm for little apparent benefit.

Why are we even doing this? Why does halting AI research feel like stopping a freight train?

I dare say it’s because we have handed over so much power to corporations.

The paperclippers are already here.

What is it with EA and AI?

PNRJ cognitive science, computer science, ethics, futurism, inequality, public policy, science and academia , , , , , , , , , ,

Jan 1 JDN 2459946

Surprisingly, most Effective Altruism (EA) leaders don’t seem to think that poverty alleviation should be our top priority. Most of them seem especially concerned about long-term existential risk, such as artificial intelligence (AI) safety and biosecurity. I’m not going to say that these things aren’t important—they certainly are important—but here are a few reasons I’m skeptical that they are really the most important the way that so many EA leaders seem to think.

1. We don’t actually know how to make much progress at them, and there’s only so much we can learn by investing heavily in basic research on them. Whereas, with poverty, the easy, obvious answer turns out empirically to be extremely effective: Give them money.

2. While it’s easy to multiply out huge numbers of potential future people in your calculations of existential risk (and this is precisely what people do when arguing that AI safety should be a top priority), this clearly isn’t actually a good way to make real-world decisions. We simply don’t know enough about the distant future of humanity to be able to make any kind of good judgments about what will or won’t increase their odds of survival. You’re basically just making up numbers. You’re taking tiny probabilities of things you know nothing about and multiplying them by ludicrously huge payoffs; it’s basically the secular rationalist equivalent of Pascal’s Wager.

2. AI and biosecurity are high-tech, futuristic topics, which seem targeted to appeal to the sensibilities of a movement that is still very dominated by intelligent, nerdy, mildly autistic, rich young White men. (Note that I say this as someone who very much fits this stereotype. I’m queer, not extremely rich and not entirely White, but otherwise, yes.) Somehow I suspect that if we asked a lot of poor Black women how important it is to slightly improve our understanding of AI versus giving money to feed children in Africa, we might get a different answer.

3. Poverty eradication is often characterized as a “short term” project, contrasted with AI safety as a “long term” project. This is (ironically) very short-sighted. Eradication of poverty isn’t just about feeding children today. It’s about making a world where those children grow up to be leaders and entrepreneurs and researchers themselves. The positive externalities of economic development are staggering. It is really not much of an exaggeration to say that fascism is a consequence of poverty and unemployment.

4. Currently the main thing that most Effective Altruism organizations say they need most is “talent”; how many millions of person-hours of talent are we leaving on the table by letting children starve or die of malaria?

5. Above all, existential risk can’t really be what’s motivating people here. The obvious solutions to AI safety and biosecurity are not being pursued, because they don’t fit with the vision that intelligent, nerdy, young White men have of how things should be. Namely: Ban them. If you truly believe that the most important thing to do right now is reduce the existential risk of AI and biotechnology, you should support a worldwide ban on research in artificial intelligence and biotechnology. You should want people to take all necessary action to attack and destroy institutions—especially for-profit corporations—that engage in this kind of research, because you believe that they are threatening to destroy the entire world and this is the most important thing, more important than saving people from starvation and disease. I think this is really the knock-down argument; when people say they think that AI safety is the most important thing but they don’t want Google and Facebook to be immediately shut down, they are either confused or lying. Honestly I think maybe Google and Facebook should be immediately shut down for AI safety reasons (as well as privacy and antitrust reasons!), and I don’t think AI safety is yet the most important thing.

Why aren’t people doing that? Because they aren’t actually trying to reduce existential risk. They just think AI and biotechnology are really interesting, fascinating topics and they want to do research on them. And I agree with that, actually—but then they need stop telling people that they’re fighting to save the world, because they obviously aren’t. If the danger were anything like what they say it is, we should be halting all research on these topics immediately, except perhaps for a very select few people who are entrusted with keeping these forbidden secrets and trying to find ways to protect us from them. This may sound radical and extreme, but it is not unprecedented: This is how we handle nuclear weapons, which are universally recognized as a global existential risk. If AI is really as dangerous as nukes, we should be regulating it like nukes. I think that in principle it could be that dangerous, and may be that dangerous someday—but it isn’t yet. And if we don’t want it to get that dangerous, we don’t need more AI researchers, we need more regulations that stop people from doing harmful AI research! If you are doing AI research and it isn’t directly involved specifically in AI safety, you aren’t saving the world—you’re one of the people dragging us closer to the cliff! Anything that could make AI smarter but doesn’t also make it safer is dangerous. And this is clearly true of the vast majority of AI research, and frankly to me seems to also be true of the vast majority of research at AI safety institutes like the Machine Intelligence Research Institute.

Seriously, look through MIRI’s research agenda: It’s mostly incredibly abstract and seems completely beside the point when it comes to preventing AI from taking control of weapons or governments. It’s all about formalizing Bayesian induction. Thanks to you, Skynet can have a formally computable approximation to logical induction! Truly we are saved. Only two of their papers, on “Corrigibility” and “AI Ethics”, actually struck me as at all relevant to making AI safer. The rest is largely abstract mathematics that is almost literally navel-gazing—it’s all about self-reference. Eliezer Yudkowsky finds self-reference fascinating and has somehow convinced an entire community that it’s the most important thing in the world. (I actually find some of it fascinating too, especially the paper on “Functional Decision Theory”, which I think gets at some deep insights into things like why we have emotions. But I don’t see how it’s going to save the world from AI.)

Don’t get me wrong: AI also has enormous potential benefits, and this is a reason we may not want to ban it. But if you really believe that there is a 10% chance that AI will wipe out humanity by 2100, then get out your pitchforks and your EMP generators, because it’s time for the Butlerian Jihad. A 10% chance of destroying all humanity is an utterly unacceptable risk for any conceivable benefit. Better that we consign ourselves to living as we did in the Neolithic than risk something like that. (And a globally-enforced ban on AI isn’t even that; it’s more like “We must live as we did in the 1950s.” How would we survive!?) If you don’t want AI banned, maybe ask yourself whether you really believe the risk is that high—or are human brains just really bad at dealing with small probabilities?

I think what’s really happening here is that we have a bunch of guys (and yes, the EA and especially AI EA-AI community is overwhelmingly male) who are really good at math and want to save the world, and have thus convinced themselves that being really good at math is how you save the world. But it isn’t. The world is much messier than that. In fact, there may not be much that most of us can do to contribute to saving the world; our best options may in fact be to donate money, vote well, and advocate for good causes.

Let me speak Bayesian for a moment: The prior probability that you—yes, you, out of all the billions of people in the world—are uniquely positioned to save it by being so smart is extremely small. It’s far more likely that the world will be saved—or doomed—by people who have power. If you are not the head of state of a large country or the CEO of a major multinational corporation, I’m sorry; you probably just aren’t in a position to save the world from AI.

But you can give some money to GiveWell, so maybe do that instead?

Is privacy dead?

PNRJ futurism, market failure, public policy , , , , , , , , , , , , , ,

May 9 JDN 2459342

It is the year 2021, and while we don’t yet have flying cars or human-level artificial intelligence, our society is in many ways quite similar to what cyberpunk fiction predicted it would be. We are constantly connected to the Internet, even linking devices in our homes to the Web when that is largely pointless or actively dangerous. Oligopolies of fewer and fewer multinational corporations that are more and more powerful have taken over most of our markets, from mass media to computer operating systems, from finance to retail.

One of the many dire predictions of cyberpunk fiction is that constant Internet connectivity will effectively destroy privacy. There is reason to think that this is in fact happening: We have televisions that listen to our conversations, webcams that can be hacked, sometimes invisibly, and the operating system that runs the majority of personal and business computers is built around constantly tracking its users.

The concentration of oligopoly power and the decline of privacy are not unconnected. It’s the oligopoly power of corporations like Microsoft and Google and Facebook that allows them to present us with absurdly long and virtually unreadable license agreements as an ultimatum: “Sign away your rights, or else you can’t use our product. And remember, we’re the only ones who make this product and it’s increasingly necessary for your basic functioning in society!” This is of course exactly as cyberpunk fiction warned us it would be.

Giving up our private information to a handful of powerful corporations would be bad enough if that information were securely held only by them. But it isn’t. There have been dozens of major data breaches of major corporations, and there will surely be many more. In an average year, several billion data records are exposed through data breaches. Each person produces many data records, so it’s difficult to say exactly how many people have had their data stolen; but it isn’t implausible to say that if you are highly active on the Internet, at least some of your data has been stolen in one breach or another. Corporations have strong incentives to collect and use your data—data brokerage is a hundred-billion-dollar industry—but very weak incentives to protect it from prying eyes. The FTC does impose fines for negligence in the event of a major data breach, but as usual the scale of the fines simply doesn’t match the scale of the corporations responsible. $575 million sounds like a lot of money, but for a corporation with $28 billion in assets it’s a slap on the wrist. It would be equivalent to fining me about $500 (about what I’d get for driving without a passenger in the carpool lane). Yeah, I’d feel that; it would be unpleasant and inconvenient. But it’s certainly not going to change my life. And typically these fines only impact shareholders, and don’t even pass through to the people who made the decisions: The man who was CEO of Equifax when it suffered its catastrophic data breach retired with a $90 million pension.

While most people seem either blissfully unaware or fatalistically resigned to its inevitability, a few people have praised the trend of reduced privacy, usually by claiming that it will result in increased transparency. Yet, ironically, a world with less privacy can actually mean a world with less transparency as well: When you don’t know what information you reveal will be stolen and misused, you will constantly endeavor to protect all your information, even things that you would normally not hesitate to reveal. When even your face and name can be used to track you, you’ll be more hesitant to reveal them. Cyberpunk fiction predicted this too: Most characters in cyberpunk stories are known by their hacker handles, not their real given names.

There is some good news, however. People are finally beginning to notice that they have been pressured into giving away their privacy rights, and demanding to get them back. The United Nations has recently passed resolutions defending digital privacy, governments have taken action against the worst privacy violations with increasing frequency, courts are ruling in favor of stricter protections, think tanks are demanding stricter regulations, and even corporate policies are beginning to change. While the major corporations all want to take your data, there are now many smaller businesses and nonprofit organizations that will sell you tools to help protect it.

This does not mean we can be complacent: The war is far from won. But it does mean that there is some hope left; we don’t simply have to surrender and accept a world where anyone with enough money can know whatever they want about anyone else. We don’t need to accept what the CEO of Sun Microsystems infamously said: “You have zero privacy anyway. Get over it.”

I think the best answer to the decline of privacy is to address the underlying incentives that make it so lucrative. Why is data brokering such a profitable industry? Because ad targeting is such a profitable industry. So profitable, indeed, that huge corporations like Facebook and Google make almost all of their money that way, and the useful services they provide to users are offered for free simply as an enticement to get them to look at more targeted advertising.

Selling advertising is hardly new—we’ve been doing it for literally millennia, as Roman gladiators were often paid to hawk products. It has been the primary source of revenue for most forms of media, from newspapers to radio stations to TV networks, since those media have existed. What has changed is that ad targeting is now a lucrative business: In the 1850s, that newspaper being sold by barking boys on the street likely had ads in it, but they were the same ads for every single reader. Now when you log in to CNN.com or nytimes.com, the ads on that page are specific only to you, based on any information that these media giants have been able to glean from your past Internet activity. If you do try to protect your online privacy with various tools, a quick-and-dirty way to check if it’s working is to see if websites give you ads for things you know you’d never buy.

In fact, I consider it a very welcome recent development that video streaming is finally a way to watch TV shows by actually paying for them instead of having someone else pay for the right to shove ads in my face. I can’t remember the last time I heard a TV ad jingle, and I’m very happy about that fact. Having to spend 15 minutes of each hour of watching TV to watch commercials may not seem so bad—in fact, many people may feel that they’d rather do that than pay the money to avoid it. But think about it this way: If it weren’t worth at least that much to the corporations buying those ads, they wouldn’t do it. And if a corporation expects to get $X from you that you wouldn’t have otherwise paid, that means they’re getting you to spend that much that you otherwise wouldn’t have—meaning that they’re getting you to buy something you didn’t need. Perhaps it’s better after all to spend that $X on getting entertainment that doesn’t try to get you to buy things you don’t need.

Indeed, I think there is an opportunity to restructure the whole Internet this way. What we need is a software company—maybe a nonprofit organization, maybe a for-profit business—that is set up to let us make micropayments for online content in lieu of having our data collected or being force-fed advertising.

How big would these payments need to be? Well, Facebook has about 2.8 billion users and takes in revenue of about $80 billion per year, so the average user would have to pay about $29 a year for the use of Facebook, Instagram, and WhatsApp. That’s about $2.50 per month, or $0.08 per day.

The New York Times is already losing its ad-supported business model; less than $400 million of its $1.8 billion revenue last year was from ads, the rest being primarily from subscriptions. But smaller media outlets have a much harder time gaining subscribers; often people just want to read a single article and aren’t willing to pay for a whole month or year of the periodical. If we could somehow charge for individual articles, how much would we have to charge? Well, a typical webpage has an ad clickthrough rate of 1%, while a typical cost-per-click rate is about $0.60, so ads on the average webpage makes its owners a whopping $0.006. That’s not even a single cent. So if this new micropayment system allowed you to pay one cent to read an article without the annoyance of ads or the pressure to buy something you don’t need, would you pay it? I would. In fact, I’d pay five cents. They could quintuple their revenue!

The main problem is that we currently don’t have an efficient way to make payments that small. Processing a credit card transaction typically costs at least $0.05, so a five-cent transaction would yield literally zero revenue for the website. I’d have to pay ten cents to give the website five, and I admit I might not always want to do that—I’d also definitely be uncomfortable with half the money going to credit card companies.

So what’s needed is software to bundle the payments at each end: In a single credit card transaction, you add say $20 of tokens to an account. Each token might be worth $0.01, or even less if we want. These tokens can then be spent at participating websites to pay for access. The websites can then collect all the tokens they’ve received over say a month, bundle them together, and sell them back to the company that originally sold them to you, for slightly less than what you paid for them. These bundled transactions could actually be quite large in many cases—thousands or millions of dollars—and thus processing fees would be a very small fraction. For smaller sites there could be a minimum amount of tokens they must collect—perhaps also $20 or so—before they can sell them back. Note that if you’ve bought $20 in tokens and you are paying $0.05 per view, you can read 400 articles before you run out of tokens and have to buy more. And they don’t all have to be from the same source, as they would with a traditional subscription; you can read articles from any outlet that participates in the token system.

There are a number of technical issues to be resolved here: How to keep the tokens secure, how to guarantee that once a user purchases access to an article they will continue to have access to it, ideally even if they clear their cache, delete all cookies, or login from another computer. I can’t literally set up this website today, and even if I could, I don’t know how I’d attract a critical mass of both users and participating websites (it’s a major network externality problem). But it seems well within the purview of what the tech industry has done in the past—indeed, it’s quite comparable to the impressive (and unsettling) infrastructure that has been laid down to support ad-targeting and data brokerage.

How would such a system help protect privacy? If micropayments for content became the dominant model of funding online content, most people wouldn’t spend much time looking at online ads, and ad targeting would be much less profitable. Data brokerage, in turn, would become less lucrative, because there would be fewer ways to use that data to make profits. With the incentives to take our data thus reduced, it would be easier to enforce regulations protecting our privacy. Those fines might actually be enough to make it no longer worth the while to take sensitive data, and corporations might stop pressuring people to give it up.

No, privacy isn’t dead. But it’s dying. If we want to save it, we have a lot of work to do.

How we sold our privacy piecemeal

PNRJ cognitive science, heuristics and biases, market failure, public policy, Uncategorized , , , , , , , , , ,

Apr 2, JDN 2457846

The US Senate just narrowly voted to remove restrictions on the sale of user information by Internet Service Providers. Right now, your ISP can basically sell your information to whomever they like without even telling you. The new rule that the Senate struck down would have required them to at least make you sign a form with some fine print on it, which you probably would sign without reading it. So in practical terms maybe it makes no difference.

…or does it? Maybe that’s really the mistake we’ve been making all along.

In cognitive science we have a concept called the just-noticeable difference (JND); it is basically what it sounds like. If you have two stimuli—two colors, say, or sounds of two different pitches—that differ by an amount smaller than the JND, people will not notice it. But if they differ by more than the JND, people will notice. (In practice it’s a bit more complicated than that, as different people have different JND thresholds and even within a person they can vary from case to case based on attention or other factors. But there’s usually a relatively narrow range of JND values, such that anything below that is noticed by no one and anything above that is noticed by almost everyone.)

The JND seems like an intuitively obvious concept—of course you can’t tell the difference between a color of 432.78 nanometers and 432.79 nanometers!—but it actually has profound implications. In particular it undermines the possibility of having truly transitive preferences. If you prefer some colors to others—which most of us do—but you have a nonzero JND in color wavelengths—as we all do—then I can do the following: Find one color you like (for concreteness, say you like blue of 475 nm), and another color you don’t (say green of 510 nm). Let you choose between the blue you like and another blue, 475.01 nm. Will you prefer one to the other? Of course not, the difference is within your JND. So now compare 475.01 nm and 475.02 nm; which do you prefer? Again, you’re indifferent. And I can go on and on this way a few thousand times, until finally I get to 510 nanometers, the green you didn’t like. I have just found a chain of your preferences that is intransitive; you said A = B = C = D… all the way down the line to X = Y = Z… but then at the end you said A > Z. Your preferences aren’t transitive, and therefore aren’t well-defined rational preferences. And you could do the same to me, so neither are mine.

Part of the reason we’ve so willingly given up our privacy in the last generation or so is our paranoid fear of terrorism, which no doubt triggers deep instincts about tribal warfare. Depressingly, the plurality of Americans think that our government has not gone far enough in its obvious overreaches of the Constitution in the name of defending us from a threat that has killed fewer Americans in my lifetime than die from car accidents each month.

But that doesn’t explain why we—and I do mean we, for I am as guilty as most—have so willingly sold our relationships to Facebook and our schedules to Google. Google isn’t promising to save me from the threat of foreign fanatics; they’re merely offering me a more convenient way to plan my activities. Why, then, am I so cavalier about entrusting them with so much personal data?

 

Well, I didn’t start by giving them my whole life. I created an email account, which I used on occasion. I tried out their calendar app and used it to remind myself when my classes were. And so on, and so forth, until now Google knows almost as much about me as I know about myself.

At each step, it didn’t feel like I was doing anything of significance; perhaps indeed it was below my JND. Each bit of information I was giving didn’t seem important, and perhaps it wasn’t. But all together, our combined information allows Google to make enormous amounts of money without charging most of its users a cent.

The process goes something like this. Imagine someone offering you a penny in exchange for telling them how many times you made left turns last week. You’d probably take it, right? Who cares how many left turns you made last week? But then they offer another penny in exchange for telling them how many miles you drove on Tuesday. And another penny for telling them the average speed you drive during the afternoon. This process continues hundreds of times, until they’ve finally given you say $5.00—and they know exactly where you live, where you work, and where most of your friends live, because all that information was encoded in the list of driving patterns you gave them, piece by piece.

Consider instead how you’d react if someone had offered, “Tell me where you live and work and I’ll give you $5.00.” You’d be pretty suspicious, wouldn’t you? What are they going to do with that information? And $5.00 really isn’t very much money. Maybe there’s a price at which you’d part with that information to a random suspicious stranger—but it’s probably at least $50 or even more like $500, not $5.00. But by asking it in 500 different questions for a penny each, they can obtain that information from you at a bargain price.

If you work out how much money Facebook and Google make from each user, it’s actually pitiful. Facebook has been increasing their revenue lately, but it’s still less than $20 per user per year. The stranger asks, “Tell me who all your friends are, where you live, where you were born, where you work, and what your political views are, and I’ll give you $20.” Do you take that deal? Apparently, we do. Polls find that most Americans are willing to exchange privacy for valuable services, often quite cheaply.

 

Of course, there isn’t actually an alternative social network that doesn’t sell data and instead just charges a subscription fee. I don’t think this is a fundamentally unfeasible business model, but it hasn’t succeeded so far, and it will have an uphill battle for two reasons.

The first is the obvious one: It would have to compete with Facebook and Google, who already have the enormous advantage of a built-in user base of hundreds of millions of people.

The second one is what this post is about: The social network based on conventional economics rather than selling people’s privacy can’t take advantage of the JND.

I suppose they could try—charge $0.01 per month at first, then after awhile raise it to $0.02, $0.03 and so on until they’re charging $2.00 per month and actually making a profit—but that would be much harder to pull off, and it would provide the least revenue when it is needed most, at the early phase when the up-front costs of establishing a network are highest. Moreover, people would still feel that; it’s a good feature of our monetary system that you can’t break money into small enough denominations to really consistently hide under the JND. But information can be broken down into very tiny pieces indeed. Much of the revenue earned by these corporate giants is actually based upon indexing the keywords of the text we write; we literally sell off our privacy word by word.

 

What should we do about this? Honestly, I’m not sure. Facebook and Google do in fact provide valuable services, without which we would be worse off. I would be willing to pay them their $20 per year, if I could ensure that they’d stop selling my secrets to advertisers. But as long as their current business model keeps working, they have little incentive to change. There is in fact a huge industry of data brokering, corporations you’ve probably never heard of that make their revenue entirely from selling your secrets.

In a rare moment of actual journalism, TIME ran an article about a year ago arguing that we need new government policy to protect us from this kind of predation of our privacy. But they had little to offer in the way of concrete proposals.

The ACLU does better: They have specific proposals for regulations that should be made to protect our information from the most harmful prying eyes. But as we can see, the current administration has no particular interest in pursuing such policies—if anything they seem to do the opposite.

The Warren Rule is a good start

PNRJ public policy , , , , , , , , , , , , , ,

JDN 2457243 EDT 10:40.

As far back as 2010, Elizabeth Warren proposed a simple regulation on the reporting of CEO compensation that was then built into Dodd-Frank—but the SEC has resisted actually applying that rule for five years; only now will it actually take effect (and by “now” I mean over the next two years). For simplicity I’ll refer to that rule as the Warren Rule, though I don’t see a lot of other people doing that (most people don’t give it a name at all).

Two things are important to understand about this rule, which both undercut its effectiveness and make all the right-wing whinging about it that much more ridiculous.

1. It doesn’t actually place any limits on CEO compensation or employee salaries; it merely requires corporations to consistently report the ratio between them. Specifically, the rule says that every publicly-traded corporation must report the ratio between the “total compensation” of their CEO and the median salary (with benefits) of their employees; wisely, it includes foreign workers (with a few minor exceptions—lobbyists fought for more but fortunately Warren stood firm), so corporations can’t simply outsource everything but management to make it look like they pay their employees more. Unfortunately, it does not include contractors, which is awful; expect to see corporations working even harder to outsource their work to “contractors” who are actually employees without benefits (not that they weren’t already). The greatest victory here will be for economists, who now will have more reliable data on CEO compensation; and for consumers, who will now find it more salient just how overpaid America’s CEOs really are.

2. While it does wisely cover “total compensation”, that isn’t actually all the money that CEOs receive for owning and operating corporations. It includes salaries, bonuses, benefits, and newly granted stock options—it does not include the value of stock options previously exercised or dividends received from stock the CEO already owns.

TIME screwed this up; they took at face value when Larry Page reported a $1 “total compensation”, which technically is true by how “total compensation” is defined; he received a $1 token salary and no new stock awards. But Larry Page has net wealth of over $38 billion; about half of that is Google stock, so even if we ignore all others, on Google’s PE ratio of about 25, Larry Page received at least $700 million in Google retained earnings alone. (In my personal favorite unit of wealth, Page receives about 3 romneys a year in retained earnings.) No, TIME, he is not the lowest-paid CEO in the world; he has simply structured his income so that it comes entirely from owning shares instead of receiving a salary. Most top CEOs do this, so be wary when it says a Fortune 500 CEO received only $2 million, and completely ignore it when it says a CEO received only $1. Probably in the former case and definitely in the latter, their real money is coming from somewhere else.

Of course, the complaints about how this is an unreasonable demand on businesses are totally absurd. Most of them keep track of all this data anyway; it’s simply a matter of porting it from one spreadsheet to another. (I also love the argument that only “idiosyncratic investors” will care; yeah, what sort of idiot would care about income inequality or be concerned how much of their investment money is going directly to line a single person’s pockets?) They aren’t complaining because it will be a large increase in bureaucracy or a serious hardship on their businesses; they’re complaining because they think it might work. Corporations are afraid that if they have to publicly admit how overpaid their CEOs are, they might actually be pressured to pay them less. I hope they’re right.

CEO pay is set in a very strange way; instead of being based on an estimate of how much they are adding to the company, a CEO’s pay is typically set as a certain margin above what the average CEO is receiving. But then as the process iterates and everyone tries to be above average, pay keeps rising, more or less indefinitely. Anyone with a basic understanding of statistics could have seen this coming, but somehow thousands of corporations didn’t—or else simply didn’t care.

Most people around the world want the CEO-to-employee pay ratio to be dramatically lower than it is. Indeed, unrealistically lower, in my view. Most countries say only 6 to 1, while Scandinavia says only 2 to 1. I want you to think about that for a moment; if the average employee at a corporation makes $50,000, people in Scandinavia think the CEO should only make $100,000, and people elsewhere think the CEO should only make $300,000? I’m honestly not sure what would happen to our economy if we made such a rule. There would be very little incentive to want to become a CEO; why bear all that fierce competition and get blamed for everything to make only twice as much as you would as an average employee?

On the other hand, most CEOs don’t actually do all that much; CEO pay is basically uncorrelated with company performance. Maybe it would be better if they weren’t paid very much, or even if we didn’t have them at all. But under our current system, capping CEO pay also caps the pay of basically everyone else; the CEO is almost always the highest-paid individual in any corporation.

I guess that’s really the problem. We need to find ways to change the overall attitude of our society that higher authority necessarily comes with higher pay; that isn’t a rational assessment of marginal productivity, it’s a recapitulation of our primate instincts for a mating hierarchy. He’s the alpha male, of course he gets all the bananas.

The president of a university should make next to nothing compared to the top scientists at that university, because the president is a useless figurehead and scientists are the foundation of universities—and human knowledge in general. Scientists are actually the one example I can think of where one individual trulycan be one million times as productive as another—though even then I don’t think that justifies paying them one million times as much.

Most corporations should be structured so that managers make moderate incomes and the highest incomes go to engineers and designers, the people who have the highest skills and do the most important work. A car company without managers seems like an interesting experiment in employee ownership. A car company without engineers seems like an oxymoron.

Finally, people who work in finance should make very low incomes, because they don’t actually do very much. Bank tellers are probably paid about what they should be; stock traders and hedge fund managers should be paid like bank tellers. (Or rather, there shouldn’t be stock traders and hedge funds as we know them; this is all pure waste. A really efficient financial system would be extremely simple, because finance actually is very simple—people who have money loan it to people who need it, and in return receive more money later. Everything else is just elaborations on that, and most of these elaborations are really designed to obscure, confuse, and manipulate.)

Oddly enough, the place where we do this best is the nation as a whole; the President of the United States would be astonishingly low-paid if we thought of him as a CEO. Only about $450,000 including expense accounts, for a “corporation” with revenue of nearly $3 trillion? (Suppose instead we gave the President 1% of tax revenue; that would be $30 billion per year. Think about how absurdly wealthy our leaders would be if we gave them stock options, and be glad that we don’t do that.)

But placing a hard cap at 2 or even 6 strikes me as unreasonable. Even during the 1950s the ratio was about 20 to 1, and it’s been rising ever since. I like Robert Reich’s proposal of a sliding scale of corporate taxes; I also wouldn’t mind a hard cap at a higher figure, like 50 or 100. Currently the average CEO makes about 350 times as much as the average employee, so even a cap of 100 would substantially reduce inequality.
A pay ratio cap could actually be a better alternative to a minimum wage, because it can adapt to market conditions. If the economy is really so bad that you must cut the pay of most of your workers, well, you’d better cut your own pay as well. If things are going well and you can afford to raise your own pay, your workers should get a share too. We never need to set some arbitrary amount as the minimum you are allowed to pay someone—but if you want to pay your employees that little, you won’t be paid very much yourself.

The biggest reason to support the Warren Rule, however, is awareness. Most people simply have no idea of how much CEOs are actually paid. When asked to estimate the ratio between CEO and employee pay, most people around the world underestimate by a full order of magnitude.

Here are some graphs from a sampling of First World countries. I used data from this paper in Perspectives on Psychological Science—the fact that it’s published in a psychology journal tells you a lot about the academic turf wars involved in cognitive economics.

The first shows the absolute amount of average worker pay (not adjusted for purchasing power) in each country. Notice how the US is actually near the bottom, despite having one of the strongest overall economies and not particularly high purchasing power:

worker_pay

The second shows the absolute amount of average CEO pay in each country; I probably don’t even need to mention how the US is completely out of proportion with every other country.

CEO_pay

And finally, the ratio of the two. One of these things is not like the other ones…

CEO_worker_ratio

So obviously the ratio in the US is far too high. But notice how even in Poland, the ratio is still 28 to 1. In order to drop to the 6 to 1 ratio that most people seem to think would be ideal, we would need to dramatically reform even the most equal nations in the world. Denmark and Norway should particularly think about whether they really believe that 2 to 1 is the proper ratio, since they are currently some of the most equal (not to mention happiest) nations in the world, but their current ratios are still 48 and 58 respectively. You can sustain a ratio that high and still have universal prosperity; every adult citizen in Norway is a millionaire in local currency. (Adjusting for purchasing power, it’s not quite as impressive; instead the guaranteed wealth of a Norwegian citizen is “only” about $100,000.)

Most of the world’s population simply has no grasp of how extreme economic inequality has become. Putting the numbers right there in people’s faces should help with this, though if the figures only need to be reported to investors that probably won’t make much difference. But hey, it’s a start.