How can we use our resources to help others the most?

This is the fundamental question of the Effective Altruism movement, and it should be the fundamental question of all charitable giving (and indeed, this post is largely copied from my similar post last year). I think the first fundamental insight of effective altruism (which really took it from Peter Singer) is that your donation can change someone’s life, and the wrong donation can accomplish nothing. People do not imagine charity in terms of “investments” and “payoffs”, yet GiveWell estimates that you can save a human life for somewhere in the magnitude of $3000.

Many American households donate that much to charity every year, and simply put, if the charities we donate to don’t try to maximize their impact, our donations may not help many people, when they could be saving a life.

This post is a short reminder that we have researched empirical evidence that you can make a difference in the world! The EA movement has already done very impressive work on how we might evaluate charitable giving, why the long term future matters, and what the most important and tractable issues might be.

Apart from the baseline incredible giving opportunities in global poverty (see GiveWell’s top charities), the long term future is an important and underfocused area of research. If humanity lives for a long time, then the vast majority of conscious humans who will exist will exist in the far future. Taking steps to ensure their existence could have massive payoffs, and concrete research in this area to avoid things like existential risk seems very important and underfunded.

I write this blog post not to shame people into donating their entire incomes (see Slate Star Codex on avoiding being eaten by consequentialist charitable impacts), but rather to ask donors to evaluate where you are sending your money within your budget and to see if perhaps the risk of paying such a high opportunity cost is worth it. Alma maters and church groups are the most common form of charity Americans give to, but the impacts from these areas seem much lower than donating to global poverty programs or the long term future.

Finally, part of this blog post is simply to publicly discuss what I donate to and to encourage others to create a charitable budget and allocate it to address problems that are large in the number of people they impact, highly neglected, and highly solvable. I thus donate about a third of my budget to GiveWell as a baseline based on evidence backed research to save lives today. I then donate another third of my budget to long term/existential risk causes where I think the impact is the highest, but the tractability is perhaps the lowest. The primary place I’ve donated to this year is the Long Term Future Fund from EA Funds. I remain uncertain on the best ways to improve the long term future, and so anything I haven’t spent from this budget item I’ve sent to GiveWell as part of my baseline giving.

The last third of my budget is reserved to focusing on policy, which is where I believe the EA movement is currently weakest. I donate money to the Center for Election Science, especially after their impressive performance this year bringing Approval Voting to St. Louis. I also donate to the Institute for Justice, as they work on fairly neglected problems in a tractable way, winning court cases to improve civil liberties for U.S. citizens. Finally, I donated a small amount to the Reason Foundation which publishes Reason magazine, as they are one of the larger places advocating big tent libertarian ideas today. It would be great to be able to move good policies to polities with bad institutions (i.e. many developing nations), but that problem seems highly intractable. It may be that the best we can do is create good institutions here and hope they are copied. I’m open to different ideas, but I am a relatively small donor and so I believe that taking risks with a portion of my donations in ways that differ from the main EA thrust is warranted.

There are many resources from the Effective Altruism community, and I’ll include several links of similar recommendations from around the EA community. If you haven’t heard of EA charities, consider giving some of your charity budget to GiveWell, or other EA organization you find convincing. If you don’t have a charity budget, consider making one for next year. Even small amounts a year can potentially save dozens of cumulative lives today, or perhaps hundreds in the far future!

Book Review: The Precipice

I have titled my annual blog post summarizing where I donate my charitable budget as “How can we use our resources to help others the most?” This is the fundamental question of the Effective Altruism movement which The Precipice‘s author, Toby Ord, helped found. For a while, Toby Ord focused on figuring out how to fight global poverty, doing the most good for the worst off people in the world. Now, he is focusing on the long term future and existential risk.

The Precipice is fantastic. It’s incredibly well written, engaging, and approachable. It covers a lot of ground from why we should care about the world, what risks humanity faces in the future, how we might think about tackling those risks, and what the future might look like if we succeed.

The Precipice eloquently interweaves fairly philosophical arguments with more empirical analysis about the sources of existential risk and tries to statistically bound them. The book discusses a pretty concerning topic of the potential end of humanity, but it does so with an eminently reasonable approach. The complexities of philosophy, science, probability, epidemiology, and more all are brought into the narrative, but made easily digestible for any reader. I honestly wish Toby Ord could teach me about everything, his writing was so clear and engaging.

The main discussion is never overwhelming with technical details, but if you ever find a point interesting, even the footnotes are amazing. At one point I came up with a counterpoint to Ord’s position, wrote that down in my notes, only to find that the next several paragraphs addressed it in its entirety, and there was actually a full appendix going into more detail. Honestly, this will be less of a book review and more of a summary with a couple final thoughts, because I think this book is not only excellent, but its content is perhaps the most important thing you can read right now. You are welcome to read the rest of this blog post, but if you have found this compelling so far, feel free to stop reading and order Toby Ord’s book posthaste.

Existential Risk

The consequences of 90% of humans on Earth dying would be pretty terrible, and given our relatively poor response to recent events, perhaps we should better explore other potential catastrophes and how we can avoid them. But The Precipice goes further. Instead of 90% of humans dying, what happens if 100% of us die out? Certainly that’s strictly worse with 100>90, but in fact these outcomes are far apart in magnitude: if all humans die out today, then all future humans never get to exist.

There’s no reason we know of that would stop our descendants from continuing to live for billions of years, eventually colonizing the stars, and allowing for the existence of trillions of beings. Whatever it is that you enjoy about humanity, whether that’s art, engineering, or the search for truth, that can’t continue if there aren’t any humans. Full stop. As far as we know, we’re the only intelligence in the universe. If we screw up and end humanity before we get off this planet, then we don’t just end it for ourselves but perhaps we end all intelligent life for the remaining trillions of years of the universe.

Even though I was aware of the broad thesis of the book, I was continually impressed with just how many different angles Ord explores. He early on notes that while we might normally think of a catastrophic extinction event, like an asteroid impact, as the thing we are keen on avoiding, in fact there are several scenarios that would be similarly devastating. For example, if humanity were to suffer some calamity that did not kill everyone but left civilization stuck at pre-industrial technology, that would also preclude humanity from living for trillions of years and colonizing the stars. A 1984 style global totalitarian state would also halt humanity’s progress, perhaps permanently.

Ord also discusses the fundamental moral philosophy implications of his thesis. The natural pitch relies on utilitarian arguments as stated above; if humanity fails to reach its potential, this not only harms any humans currently alive but all future generations. Other arguments against extinction include a duty to our past and what we owe to our ancestors, the rights of those future generations who don’t get to decide for themselves, and the simple fact that we would lose everything we currently value.

The book categorizes three types of risk: natural, anthropogenic, and future risks. Natural includes asteroids, supervolcanoes, and stellar explosions. These are pretty diverse topics, and Ord is quite informative. The story about asteroid risk was particularly fascinating to me. In the 90s, the relatively new discovery of the dinosaurs’ demise led Congress to task NASA with identifying all the largest near-Earth asteroids to see if they pose a threat to Earth. They allocated some money, and NASA tracked every near-Earth asteroid over 10 km in length, and determined that none pose a threat in the next century. They then moved on to 1 km asteroids and have now mapped the vast majority of those as well. The total cost of the program was also quite small for the information provided — only $70 million.

This is one of the rare successes in existential risk so far. Unfortunately, as Ord points out several times in the book, current foundational existential risk research at present is no more than $50 million a year. Given the stakes, this is deeply troubling. As context, Ord points out that the global ice cream market is about $60 billion, some 1000x larger.

I’ll skip the other natural risks here, but the book bounds natural risk quite skillfully; humans have been around for about 200,000 years, so it seems natural risk can’t be much higher than 0.05% per century. Even then, we’d expect our technologically advanced civilization to be more robust to these risks than we have been in the past. Many species survived even the largest mass extinctions, and none of them had integrated circuits, written language, or the scientific method.

That doesn’t mean that all risk has declined over time. On the contrary, according to Ord, the vast majority of existential risk is anthropogenic in origin. Nuclear weapons and climate change dominate this next section. It’s remarkable just how callous early tests of nuclear weapons really were. Ord recounts how there were two major calculations undertaken by a committee of Berkeley physicists before the Manhattan project got underway in earnest. One was whether the temperature of a sustained nuclear reaction would ignite the entire atmosphere in a conflagration (the committee believed it would not). The other was whether Lithium-7 would contribute to a thermonuclear explosion (it was believed it would not). It turns out that Lithium-7 can contribute to a thermonuclear explosion as was found out when the Castle Bravo test was about three times larger than expected, irradiating some 15 nearby islands.

It turned out the other calculation was correct, and the first nuclear explosion in 1945 did not ignite the atmosphere. But clearly, given the failure of the other calculation, the level of confidence here was not high enough to warrant the risk of ending all life on Earth.

Luckily, current risk from nuclear weapons and climate change that would wipe out humanity seems quite low (although not zero). Even a nuclear winter scenario or high sea level rise would not make the entire Earth uninhabitable, and it is likely humans could adapt, although the loss of life would still be quite catastrophic.

Instead, the bulk of the risk identified by Toby Ord is in future technologies which grow more capable every year. These include engineered pandemics from our increasingly powerful and cheap control over DNA synthesis, as well as artificial intelligence from our increasingly powerful and integrated computer systems.

The threat of engineered pandemics is particularly prescient as I write this in August 2020 where SARS-CoV-2 is still sweeping the world. Ord notes that even given quite positive assumptions about whether anyone would want to destroy the world with a virus, if the cost is cheap enough, it only takes one crazy death cult to pull the trigger. Even an accidental creation of a superweapon is a serious risk, as production is cheap and there are many examples of accidental leakages of bioweapons from government laboratories in the past. Unfortunately, we are also woefully unprepared on this front. The Biological Weapons Convention had a budget of $1.4 million in 2019, which Ord notes is less than most McDonald’s franchises.

Risks from unaligned artificial intelligence are similarly related to technical advancements. Ord notes that artificial intelligence has had some impressive achievements recently from photo and face identification to translation and language processing to games like Go and Starcraft. As computer hardware gets better and more specific, and as we discover more efficient algorithmic applications of artificial intelligence, we should expect this trend to continue. It therefore seems plausible that sometime in the future, perhaps this century, we will see artificial intelligence exceed human ability in a wide variety of tasks and ability. The Precipice notes that, were this to happen with some sort of general intelligence, humanity would no longer be the most intelligent species on the planet. Unless we have some foresight and strategies in place, having a superior intelligence with it own goals could be considerably dangerous.

Unfortunately, we are already quite poor at getting complex algorithms to achieve complicated goals without causing harm (just take a look at the controversy around social media and misinformation, or social media and copyright algorithms). The use of deep learning neural networks in more high stakes environments means we could be facing opaque algorithmic outcomes from artificial intelligence that we don’t know if we’ve correctly programmed to achieve the goals we actually want. Throw in the fact that human civilizational goals are multifaceted and highly debated, and there is a great deal of mistakes that could occur between what humans “want” and what a superior intelligence attempts to accomplish. While Toby Ord doesn’t think we should shut down AI research, he does suggest we take this source of risk more seriously by devoting resources to addressing it and working on the problem.

So What Do We Do?

I’ve spent a lot of time on enumerating risks because I think they are a concrete way to get someone who is unfamiliar with existential risk to think about these ideas. But Ord isn’t writing a book of alarmism just to freak out his audience. Instead, starting with the high levels of risk and adding the extremely negative consequences, Ord details how we might begin to tackle these problems. Unprecedented risks come with modeling challenges: if an existential risk cannot by definition, have ever occurred, how can we know how likely it is? We have to acknowledge this limitation, use what incomplete knowledge we can have access to (number of near misses is a good start), and start building institutions to focus on solving these hard problems.

International coordination is a major factor here. Many of these problems are collective action problems. Humanity has found ways around collective action issues with international institutions before (nuclear arms treaties), and so we need to replicate those successes. Of course, we can’t establish new or better institutions unless we get broad agreement that these issues are major problems that need to be solved. Obviously, that’s why Ord wrote this book, but it’s also why I feel compelled to blog about it as well. More on that momentarily.

In this section of the book, The Precipice outlines preliminary directions we can work towards to improve our chances of avoiding existential catastrophes. These include obvious things like increasing the funding for the Biological Weapons Convention, but also discussions on how to think about technological progress, since much of our future existential risk rises as technology improves. We also obviously need more research on existential risk generally.

Finally, I want to wrap up discussing Appendix F, which is all of Ord’s general policy recommendations put into one place. As policy prioritization has long been an interest of mine, I found Toby Ord’s answer to be quite fascinating. I wrote a post a few months back discussing the highest impact policies actually being discussed in American politics in this election cycle. Comparing it to Toby Ord’s recommendations, the overlap is essentially nonexistent except for some points on climate change, which most democrats support such as the U.S. rejoining the Paris Climate Agreement. There’s also a point about leveraging the WHO to better respond to pandemics, and given Trump has essentially done the exact opposite by removing U.S. funding for the WHO, I suppose I should at least include that as relevant policy debate.

I want to emphasize that Ord has 9 pages of policy ideas, and many of them are likely uncontroversial (improve our understanding of long period comets, have the Biological Weapons Convention have a real budget), but our political system is failing to even address these challenges, and I think it’s important to highlight that.

There is room for optimism; human knowledge is improved by discussion and research, and that includes reading and blogging. If you find these ideas interesting, or even more broadly, if you think there are valuable things in the world, one of the most effective activities you could do this year might be to just read The Precipice. Even without the weight of humanity, the concepts, problem solving, and prose are worth the read all by themselves. This is definitely by favorite book I’ve read this year, and I’ve skipped over summarizing whole sections in the interests of time. Ord even has a whole uplifting chapter about humanity’s future potential, and is overall quite positive. Please attribute any gloominess on this topic to me and not the book.

And if you do read this book, it also just makes for intriguing conversation. I couldn’t help but tell people about some of the ideas here (“are supervolcanoes a national security threat?” ), and the approach is just wonderfully different, novel, and cross-disciplinary.

For more on this, but slightly short of reading the whole book, I also recommend Toby Ord’s excellent interview on the 80000 Hours Podcast. On that page you can also find a host of awesome links to related research and ideas about existential risk. I’ll also link Slate Star Codex’s longer review of The Precipice, and places to buy it.

How can we use our resources to help others the most?

This is the fundamental question of the Effective Altruism movement, and it should be the fundamental question of all charitable giving. I think the first fundamental insight of effective altruism (which really took it from Peter Singer) is that your donation can change someone’s life, and the wrong donation can accomplish nothing. People do not imagine charity in terms of “investments” and “payoffs”, yet GiveWell estimates that you can save a human life for somewhere in the magnitude of $2500.

Many American households donate that much to charity every year, and simply put, if the charities we donate to don’t try to maximize their impact, our donations may not help many people, when they could be saving a life.

This post is a short reminder that we have researched empirical evidence that you can make a difference in the world! The EA movement has already done very impressive work on how we might evaluate charitable giving, why the long term future matters, and what the most important and tractable issues might be.

Apart from the baseline incredible giving opportunities in global poverty (see GiveWell’s top charities), the long term future is an important and underfocused area of research. If humanity lives for a long time, then the vast majority of conscious humans who will exist will exist in the far future. Taking steps to ensure their existence could have massive payoffs, and concrete research in this area to avoid things like existential risk seems very important and underfunded.

I write this blog post not to shame people into donating their entire incomes (see Slate Star Codex on avoiding being eaten by consequentialist charitable impacts), but rather to ask donors to evaluate where you are sending your money within your budget and to see if perhaps the risk of paying such a high opportunity cost is worth it. Alma maters and church groups are the most common form of charity Americans give to, but the impacts from these areas seem much lower than donating to global poverty programs or the long term future.

Finally, part of this blog post is simply to publicly discuss what I donate to and to encourage others to create a charitable budget and allocate it to address problems that are large in the number of people they impact, highly neglected, and highly solvable. I thus donate about a third of my budget to GiveWell as a baseline based on evidence backed research to save lives today. I then donate another third of my budget to long term causes where I think the impact is the highest, but the tractability is perhaps the lowest. Top charities I’ve donated to here include the Machine Intelligence Research Institute for AI alignment research, as well as the Long Term Future Fund from EA Funds.

The last third of my budget is reserved to focusing on policy, which is where I believe the EA movement is currently weakest. I donate money to the Institute for Justice, as they work on fairly neglected problems in a tractable way, winning court cases to improve civil liberties for U.S. citizens. I also like the Center for Election Science as they work to improve the democratic processes in the US. It would be great to be able to move good policies to polities with bad institutions (i.e. many developing nations), but that problem seems highly intractable. It may be that the best we can do is create good institutions here and hope they are copied. I’m open to different ideas, but I am a relatively small donor and so I believe that taking risks with a portion of my donations in ways that differ from the main EA thrust is warranted. This is by far my most uncertain category, and thus usually I will not entirely fulfill my budget for policy charities. I plan on giving anything remaining to GiveWell.

There are many resources from the Effective Altruism community, and I’ll include several links of similar recommendations from around the EA community. If you haven’t heard of EA charities, consider giving some of your charity budget to GiveWell, or other EA organization you find convincing. If you don’t have a charity budget, consider making one for next year. Even small amounts a year can potentially save dozens of cumulative lives!

Artificial General Intelligence and Existential Risk

The purpose of this post is to discuss existential risk, and why artificial intelligence is a relatively important aspect of existential risk to consider. There are other essays about the dangers of artificial intelligence that I will link to throughout and at the end of this post. This essay is a different approach that perhaps will appeal to someone who has not seriously considered artificial general intelligence as an issue requiring civilization’s attention. At the very least, I’d like to signal that it should be more socially acceptable to discuss this problem.

First is the section on how I approached thinking about existential risk. My train of thought is a follow up to Efficient Advocacy. Also worth reading: Electoral Reform Fantasies.

Background

Political fights, especially culture war battles that President Trump seems so fond of, are loud, obnoxious, and tend to overshadow more impactful policy debates. For example, abortion debates are pretty common, highly discussed political issues, but there have been almost no major policy changes since the Supreme Court’s decision 40 years ago.  The number of abortions in the US has declined since the 1980s, but it seems uncorrelated with any political movements or electoral victories. If there aren’t meaningful differences from different political outcomes, and if political effort, labor, and capital is limited, these debates seem to distract from other areas that could impact more people. Trump seems especially good at finding meaningless conflicts to divide people, like NFL players’ actions during the national anthem or tweeting about Lavar Ball’s son being arrested in China.

Theorizing about how to combat this problem, I started making a list of what might be impactful-but-popular (or at least not unpopular) policies that would make up an idealized congressional agenda: nominal GDP futures markets, ending federal prohibition of marijuana, upgrading Social Security Numbers to be more secure, reforming bail. However, there is a big difference between “not unpopular”, “popular”, and “prioritized”. I’m pretty sure nominal GDP futures markets would have a pretty positive effect on Federal Reserve policy, and I can’t think of any political opposition to it, but almost no one is talking about it. Marijuana legalization is pretty popular across most voters, but it’s not a priority, especially for this congress. So what do you focus on? Educating more people about nominal GDP futures markets so they know such a solution exists? Convincing more people to prioritize marijuana legalization?

The nagging problem is that effective altruist groups like GiveWell have taken a research based approach to identify at what the best ways are to use our money and time to improve the world. For example, the cost of distributing anti-mosquito bed nets is extremely low, resulting in an average life saved from malaria at a cost in the thousands of dollars. The result is that we now know our actions have a significant opportunity cost; if a few thousand dollars worth of work or donations doesn’t obviously have as good an impact as literally saving someone’s life, we need a really good argument as to why we should do that activity as opposed to contributing to GiveWell’s top charities.

One way to make a case as to why there are other things worth spending money on besides GiveWell’s top charities, is to take a long term outlook, trying to effect a large change that would impact a large amount of people in the future.  For example, improving institutions in various developing countries would help those populations become richer. Another approach would be to improve the global economy, which would both allow for more investment in technology as well as push investment into developing countries looking for returns. Certainly long term approaches are more risky compared to direct impact charities that improve outcomes as soon as possible, but long term approaches can’t be abandoned either.

Existential Risk

So what about the extreme long term? What about existential risk? This blog’s philosophy takes consequentialism as a founding principle, and if you’re interested in the preceding questions of what policies are the most helpful, and where we should focus our efforts, you’ve already accepted that we should be concerned about the effects of our actions. The worst possible event, from a utilitarian perspective would be the extinction of the human race, as it would not just kill all the humans alive today (making it worse than a catastrophe that only kills half the humans), but also ends the potential descendants of all of humanity, possibly trillions of beings. If we have any concern for the the outcomes of our civilization, we must investigate sources of existential risk. Another way to state this is: assume it’s the year 2300, and humans no longer exist in the universe. What is the most likely cause of our destruction?

Wikipedia actually has a very good article on Global Catastrophic Risk, which is a broad category encompassing things that could seriously harm humanity on a global scale. Existential risks are a strict subset of those events, which could end humanity’s existence permanently. Wikipedia splits them up into natural and anthropogenic. First, let’s review the non-anthropogenic risks (natural climate change, megatsunamis, asteroid impacts, cosmic events, volcanism, extraterrestrial invasion, global pandemic) and see whether they qualify as existential.

Natural climate change and megatsunamis do not appear to be existential in nature. A megatsunami would be terrible for everyone living around the affected ocean, but humans on the other side of the earth would appear to be fine. Humans can also live in a variety of climates, so natural climate change would likely be slow enough for some humans to adapt, even if such an event causes increased geopolitical tensions.

Previous asteroid impacts have had very devastating impacts on Earth, notably the Cretaceous-Paleocene extinction event some 66 million years ago. This is a clear existential risk, but you need a pretty large asteroid to hit Earth, which is unusual. Larger asteroids can also be more easily identified from further away, giving humanity more time to do something (push it off path, blow it up, etc). The chances here are thus pretty low.

Other cosmic events are also low probability. Gamma-ray bursts are pretty devastating, but they’d have to be close-by (with a few hundred light-years at least) as well as aimed directly at Earth. Neither of these is likely within the next million years.

Volcanism is also something that has the potential to be pretty bad, perhaps existential level (see Toba Catastrophe Theory), but it is also pretty rare.

An alien invasion could easily destroy all of humanity. Any species with the capability to travel across interstellar space with military ambitions would mean they are extremely technologically superior. However, we don’t see any evidence of a galactic alien civilization (see Fermi Paradox 1 & 2 and The Great Filter). Additionally, solving this problem seems somewhat intractable; on a cosmic timescale, an alien civilization that arose before our own would likely have preceded us by millennia, meaning the technology gap between us and them would be hopelessly and permanently large.

A global pandemic seems pretty bad, certainly much more likely than anything else we’ve covered in the short term. This is also exacerbated by human actions creating a more interconnected globe. However, it is counterbalanced by the fact that no previous pandemic has ever been 100% lethal, and that modern medicine is much better than it was during the Black Plague. This is a big risk, but it may not be existential. Definitely on our shortlist of things-to-worry-about though.

Let’s talk about anthropogenic risks next: nuclear war, conventional war, anthropogenic climate change, agricultural crises, mineral exhaustion, artificial intelligence, nanotechnology, biotechnology.

A common worry is nuclear war. A massive nuclear exchange seems somewhat unlikely today, even if a regional disagreement in the Korean peninsula goes poorly in the worst possible way. It’s not common knowledge, but the “nuclear winter” scenario is still somewhat controversial, and I remain unconvinced that it poses a serious existential threat, although clearly a nuclear exchange would kill millions. Conventional war is also out as it seems strictly less dangerous than a nuclear war.

For similar reasons to nuclear winter, I’m not quite worried about global warming on purely existential terms. Global warming may be very expensive, it may cause widespread weather, climate, and ecological problems, but I don’t believe humanity will be entirely wiped out. I am open to corrections on this.

Agricultural crises and mineral exhaustion seem pretty catastrophic-but-not-existential as well. These would result in economic crises, but by definition, economic crises need humans to exist; if there are fewer humans, it seems that an agricultural crisis would no longer be an issue.

The remaining issues are largely technological in nature: artificial intelligence, biotechnology, nanotechnology, or technical experiments going wrong (like if the first nuclear test set the atmosphere on fire). These all seem fairly concerning.

Technological Existential Risk

Concern arises because technological progress means the likelihood that we will have these technologies grows over time, and, once they exist, we would expect their cost to decrease. Additionally, unlike other topics listed here, these could wipe out humanity permanently. For example, a bioengineered virus could be far more deadly than what would naturally occur, possibly resulting in a zero survival rate. The cost of DNA technology has steadily dropped, and so over time, we might expect the number of organizations or people who have the knowledge and funding to engineer deadly pathogens to increase. The more people who have this ability, the more likely that someone makes a mistake and releases a deadly virus that kills everyone. An additional issue is that it is quite likely that military research teams are right now researching bioweapons like an engineered pathogen. Incentives leading to the research of dangerous weapons like this are unlikely to change, even as DNA engineering improves, meaning the risk of this threat should grow over time.

Nanotechnology also has the potential to end all life on the planet, especially under a so-called “grey goo” scenario, where nanobots transform all the matter on Earth. This has a lot of similarities to a engineered pathogen, except the odds of any human developing some immunity no longer matter, and additionally all non-human life, indeed, all matter on Earth is also forfeit, not just the humans. Like biotechnology threats, we don’t have this technology yet, but it is an active area of research. We would also expect this risk to grow over time.

Artificial General Intelligence

Finally, artificial general intelligence contains some similar issues to the others: as technology advances, we have a higher chance of creating it; the more people who can create it, the more dangerous it is; once it is created, it could be deadly.

This post isn’t a thesis on why AI is or isn’t going to kill all humans. We made an assumption that we were looking exclusively at existential risk in the near future of humanity. Given that assumption, our question is why will AI be more likely to end humanity than anything else? Nonetheless, there are lingering questions as to whether AI is an actual “real” threat to humanity, or just an unrealistic sci-fi trope. I will outline three basic objections to AI being dangerous with three basic counterarguments.

The first objection is that AI itself will not be dangerous because it will be too stupid. Related points are that AI is too hard to create, or we can just unplug it if it has differing values from us. Counterarguments are that experts disagree on exactly when we can create human-level AI, but most agree that it’s plausible in the next hundred or couple hundred years (AI Timelines). It’s also true that we’ve seen improvements in AI ability to solve more general and more complex problems over time; AlphaZero learned how to play both Go and Chess better than any human without changes in its base code, YouTube uses algorithms to determine what content to recommend and what content to remove ads from, scanning through thousands of hours of video content every minute, Google’s Pixel phone can create software based portrait photos via machine learning rather than needing multiple lenses. We should expect this trend to continue, just like with other technologies.

However, the difference between other technological global risks and AI is that the machine learning optimization algorithms could eventually be applied to machine learning itself. This is the concept of an “intelligence explosion”, where an AI uses its intelligence to design and create successively better versions of itself. Thus, it’s not just that an organization might make a dangerous technological breakthrough, like an engineered virus, but that once the breakthrough occurs, the AI would rapidly become uncontrollable and vastly more intelligent than us. The intelligence analogy being that a mouse isn’t just less smart than a human, it literally doesn’t comprehend that its environment can be so manipulated by humans that entire species depend on the actions of humans (i.e. conservation, rules about overhunting) for their own survival.

Another objection is that if an AI is actually as intelligent as we fear it could be, it wouldn’t make “stupid” mistakes like destroying all of humanity or consuming the planet’s resources, because that wouldn’t count as “intelligent”. The counterpoint is the Orthogonality Thesis. This simply states that an AI can have any goal. Intelligence and goals are orthogonal and independent. Moreover, an AI’s goal does not have to explicitly target humans as bad (e.g. “kill all the humans”) to cause us harm. For example, a goal to calculate all the digits of pi or solve the Riemann Hypothesis might require as much computing power as possible. As part of achieving this goal, a superintelligence would determine that it must manufacture computing equipment and maximize energy to its computation equipment. Humans use energy and are made of matter, so as a way to achieve its goal, it would likely exterminate humanity, and convert all matter it could into computation equipment. Due to its superintelligence, it would accomplish this.

A final objection is that despite experts believing human level AI will happen in the next 100 years, if not sooner, there is nothing to be done about it today or that it is a waste of time to work on this problem now. This is also known as the “worrying about overpopulation on Mars” objection, comparing the worry about AI to something that is several scientific advancements away.  Scott Alexander has an entire blog post on this subject, which I recommend checking out. The basic summary is that AI advancement and AI alignment research are somewhat independent. And we really need to learn how to properly align AI values before we get human level AI.

We have a lot of theoretical philosophy that we need to figure out how to impart to a computer. Things like how humans actually make decisions, or how to value different moral tradeoffs. This could be extraordinarily complicated, as an extremely smart optimization algorithm could misinterpret almost everything we say if it did not already share our values for human life, health, and general brain state. Computer scientists set out to teach computers how to understand natural human language some 60 years ago, and we still haven’t quite nailed it. If imparting philosophical truths is similarly difficult, there is plenty of work to be done today.

Artificial intelligence could advance rapidly from human level to greater than human very quickly; the best human Go player lost to an AI (AlphaGo) in 2016, and a year later, AlphaGo lost to a new version, AlphaGo Zero, 100 games to none. It would thus not be surprising if a general intelligence achieved superhuman status a year after achieving human-comparable status, or sooner. There’s no fire alarm for artificial general intelligence. We need to be working on these problems as soon as possible.

I’d argue then, that of all scenarios listed here, a misaligned AI is the most likely to actually destroy all of humanity as a result of the Orthogonality Thesis. I also think that unlike many of the other scenarios listed here, human level AI will exist sometime soon, compared to the timescale of asteroids and vulcanism (see AI Timelines, estimates are highly variable, anywhere from 10 to 200 years). There is also a wealth of work to be done surrounding AI value alignment. Correctly aligning future AI with goals compatible with human values is thus one of the most important challenges facing our civilization within the next hundred years or so, and probably the most important existential threat we face.

The good news is that there are some places doing this work, notably the Machine Intelligence Research Institute, OpenAI, and the Future of Humanity Institute. The bad news is that despite the importance of this issue, there is very little in the way of conversations, money, or advocacy. AI Safety research is hard to calculate in total, as some research is likely done by private software companies, but is optimistically on the order of tens of millions of dollars a year. By comparison, the U.S. Transportation Security Administration, which failed to find 95% of test weapons in a recent audit, costs $7.5 billion a year.

Further Reading

I have focused this essay on trying to convey the mindset of thinking about existential risk generally and why AI is specifically worrying in this context. I’ve also tried to keep it short. The following are further resources on the specifics of why Artificial General Intelligence is worth worrying about in a broader context, arranged by length. If you felt my piece did not go in depth enough on whether AI itself is worth being concerned about, I would urge you to read one of the more in depth essays here which focus on that question directly.

 


Leave a comment on the official reddit thread.