TL;DR: Many themes of the Jurassic Park book have unique parallels to the rise of Artificial Intelligence. Relaxed regulations, rapid development, and an incomplete understanding of the product are just a few of the parallel themes.
I recently re-read Jurassic Park by Micheal Crichton (can you believe it was a book before it was a movie and a reboot?). If you are unfamiliar with the book/movies, they tell the story of an older businessman, John Hammond, with an unrelenting desire to be the first to develop the newest form of entertainment that is the product of quikly developed private research, namely, the ability to re-create dinosaurs. The two sentence summary of the whole book is that John Hammond has some scientists come check out his Zoo for dinasours to appease his investors and make sure everything is safe. SPOILERS Things go terribly and people die.
As I read the book many themes started to present themselves, including - don’t rush your development and know your product. While the story is fiction, I was surprised at how closely it can be compared to what is happening in Silicon valley and AI. The philosophies of John Hammond, the main person responsible for pushing the research behind how dinasours came back to life, almost directly relate to the philosophies that silicon valley companies have towards AI.
Wow, what a dumb comparison you might think, but give me the chance to convince you. I think if you’ve read the book recently, you probably don’t think that. Let me explain a couple of reasons why I think the two are similar.
1. Genetically Engineering Dinosaurs and simulating human intelligence are arguably similar.
I can’t use a quote from the book to make this comparison, but the idea is still valid. DNA is a double helix with hundreds of millions of pairs of nucleotides. These pairs encode genes, which are then expressed to create certain traits of an animal/person whatever. What sequences within the hundreds of millions of pairs are responsible for expressing the genes is not yet an exact science. Comparitively, AI aims to encode a bunch of words as vectors in a multi-dimensional space to provide similarities between words. AI also models the neurons in the brain, if you provide these simulated neurons with a question that is just many multidimensional vectors, you can get a response is another set of multi-dimensional vectors. Looking those back up in your vector encoded dictionary you can create a human understandable response. Both DNA/gene encoding and AI development, are still under active development, with experiments having side effectst that are still not yet completely understood. Sometimes weird traits are expressed from a DNA strand, but it’s still not completely understood what protein within the DNA is responsible for expression that trait. Sometimes AI hallucinates, and we still don’t know why. Both of these areas of science are still actively being developed.
2. Regulating a Zoo of Genetically Engineered Dinosaurs has Similarities to Regulating AI.
“But InGen was obviously setting up one of the most powerful genetic engineering facilities in the world in an obscure Central American country. A country with no regulations.”(page 43)
The book doesn’t make it clear if Hammond chose the island for that exact reason, but the comparison can still be made. There is little to no regulation on an island hundreds of miles off the coast of Costa Rica. Currently there are no regulations around AI (at least in the U.S), and no plans to make any.
The following passage also stood out to me. AI is not seen as research to save lives, but more as a tool to improve quality of life/productivity. For this reason, it becomes difficult to convince the public that regulation will be needed.
“The original genetic engineering companies, like Genentech and Cetus, were all starting to make pharmaceuticals. New drugs for mankind. Noble, noble prupose. Unfortunately, drugs face all kinds of barriers. FDA testing alone takes five to eight years, if you’re lucky. Even worse, there are forces at work in the marketplace. Suppose you make a miracle drug for cancer or heart disease – as Genentech did. Suppose you now want to charge a thousand dollars tor two thousand dollars a dose. You might imagine that is your privilege. After all, you invested the drug, you paid to develop and test it; you should be able to charge whatever you wish. But do you really think that the government will let you do that? No, Henry, they will not. Sick people aren’t going to pay a thousand dollars a dose for needed medication–they won’t be grateful, they’ll be outraged. Blue Cross isn’t going to pay it. … So something will happen. Your patent application will be denied. Your permits will be delayed. Something will force you to see reason and to sell your drug at a lower cost.”(page 222-223)
3. We do not fully understand how AI behaves, just like how Jurassic Park administrators did not know the behavior of Dinosaurs.
We are still learning how AI behaves and responds to certain inputs, and there is literally nothing we can directly compare AI with. In the book, they try to have zookeepers from big game reserves tending to the dinosaurs with the idea that they know how to care for larger animals. However, the zookeepers don’t know how to take care of dinosaurs. This lack of knowledge of how dinosaurs behave would end up leading to multiple issues within the park.
“And, finally, we have the unprecendented problems of caring for a population of animals that no one has ever tried to maintain before.” Oh, it’s not as bad as all that,” Hammond said. “Yes, it is. Your’re just not here to see it,” Arnold said. “The tyrannosaurs drink the lagoon water and sometimes get sick; we aren’t sure why. The triceratops femailes ill one another in fights for dominance and have to be separated into groups smaller than six. We don’t know why. The stegosaurs frequently get blisters on their tongues and diarrhea, for reasons no one yet understands, even though we’ve lost two. Hypsilophodonts get skin rashses. And the velociraptors–” (page 156)
Jurassic Park still had many problems, without knowing know why or what the fix should be. With AI, as mentioned earlier, scientists are still understanding the behavior for certain AI models, but I’m sure there are still a number of smaller issues that cannot be explained yet, and they are still hidden from consumers.
4. The abilites of AI are rapidly growing, with newer better versions coming out every week. John Hammond rushes the development of the dinosaurs and the park so it can open ASAP. The lead researcher, Henry Wu, and the zookeeper, John Arnold want to wait and develop less aggressive dinosaurs.
"We’ve done everything we can to promote growth, and accelerate development into adulthood." Hammond shrugged. "That was inevitable. We didn’t want to wait. We have investors to consider." "Of course. But I’m just saying, why stop there? Why not push ahead to make exactly the kind of dinaosaur the we’d like to see? Once that is more acceptable to visitors, and one that is easier for us to handle? A slower, more docile version for our park?"(page 137)
I would bet a lot of money that this kind of conversation happens at tech companies with investors, just not about dinosaurs obviously. It’s clear though that the LEAD RESEARCHER doesn’t think that the current PRODUCTS/dinosaurs are safe for the GENERAL PUBLIC, and that more time needs to be spent developing a safer version for consumption. The CEO, however, wants to appease the investors/shareholders. I imagine these conversations have happened with AI product managers & company leadership, except the situatino today is worse. The investors/shareholders for Microsoft, OpenAI, Google, and Apple are competing with each other to develop the next best version, but are probably not listening to the researchers about the ethical dilemmas, or how it could affect public safety. The companies are only thinking about their shareholders/investors.
Later on in the book there is an argument between Hammond and John Arnold the chief systems engineer for Jurassic Park. (page 155-156)
“You’ve got to realize that, from an engineering standpoint, Jurassic Park is by far the most ambitious theme park in history… Jurassic Park has all the problems of any amusement park, ride maintenance, queue control, transportation, food handling, living accommodations, trash disposal, security. Second we have all the problems of a major zoo – care of the animals; health and welfare; feeding and cleanliness protection from insects, pests, allergies, and illnesses; maintenance of barrieres; and all the rest. And finally, we have the unprecedented problems of caring for a population of animals that no one has ever tried to maintain before.”
Arnold, the zookeeper, then goes on to list a few of the problems they are having with the animals, and consistently mentions that no one knows what the underlying issue is, and they have no idea where to start. These conversations with Hammond show that a lot of the time, the main investors don’t grasp what issues there may be and how they could put their own customers in danger. I fear this could be happening with AI, but the customers don’t know.
5. John Hammond has researchers recreating life by developing dinosaurs. A completely new life form that no one has experience with. Top researchers don’t quite understand how dinosaurs behave, so neither will the general public. The following quote is from Ian Malcom, one of the researchers on the island.
“I will tell you what I am talking about … Most kinds of power require substantial sacrifice by whoever wants the power. There is an apprenticeship, a discipline lasting many years. Whatever kind of power you want. President of the company. Black belt in karate. Spiritual guru. Whatever it is you seek, you have to put in the time, the practice, the effort. You must give up a lot to get it. It has to be very important to you. And once you have attained it, it is your power. It can’t be given away: it resides in you. It is literally the result of your discipline.”
“Now, what is interesting about this process is that, by the time someone has acquired the ability to kill with his bare hands, he has also matured to the point where he won’t use it unwisely. So that kind of power has a built-in control. The discipline of getting the power changes you so that you won’t abuse it.”
“But scientific power is like inherited wealth: attained without discipline. You read whatever others have done, and you take the next step. You can do it very young. You can make progress very fast. There is no discipline lasting many decades. There is no mastery: old scientists are ignored. There is no humility before nature. There is only a get-rich-quick make-a-name-for-yourself-fast philosphy. Cheat, lie, falsify–it doesn’t matter. Not to you, or to your colleagues. No one will criticize you. No one has any standards. They are all trying to do the same thing: to do something big, and do it fast.”
“And because you can stand on the shoulders of giants, you can accomplish something quickly. You don’t even know exactly what you have done, but already you have reported it, patented it, and sold it. And the buyer will have even less discipline than you. The buyer simply purchases the power, like any commodity. The buyer doesn’t even conceive that any discipline might be necessary.”
…
I’ll make it simple, … A karate master does not kill people with his bare hnads. He does not lose his temper and kill his wife. The person who kills is the person who has no discipline, no restraint, and who has purchased his power in the form of a Saturday night special. And that is the kind of power that science fosters, and permits. And that is why you thinkg that to build a place like this is simple. (page 343-344).
AI is being developed to be sold to the masses, but the masses won’t know how to use it properly. We are introducing what some may consider independent life forms into an environment we cannot control, and companies want to sell it to the general public. We already have seen multiple instances of students using it to cheat on assignments. We have also seen AI hallucinate results for research papers. The people who have developed AI know the best use cases, but when anybody can buy a subscription, they don’t use it wisely. They could end up killing someone because they saw a saturday night karate special.
Ethical Reasons
There are also many ethical concerns and/or reasons for why Jurassic Park can be compared to AI development. I will save those for another blog post when I eventually read Jurassic Park again.
Conclusion
Jurassic Park can of course be compared to more scientific discoveries that eventually made their way towards general consumption. Think airplanes, the Iphone, cars, personal computers etc. The problem with AI and dinosaurs, however, is that they can take general action without user input. AI and dinosaurs can make decisions that cannot always be explained by humans. The other products mentioned earlier would only act given user provided input. A car won’t drive unless the user presses on the gas, A plane won’t fly unless a pilot instructs it to. The behaviour can always be predicted via the laws of physics and user input. AI, however, eventually, will be be developed to make decisions and behave without specific user input. We will not be able to predict every AI action, and whatever action it takes can then affect the broader world without knowledge from any users. This will be very dangerous in my opinion.