Once a week.
Subscribe to our weekly newsletter.
The First Track of the First Album Composed and Produced by AI
The first pop album composed and produced by AI, and Taryn Southern.
At Big Think, we’ve been intrigued for a while about the intersection between artificial intelligence/machine learning and creativity. There was a crazy AI-written movie script a while back, and a couple of attempts by AI at producing music . In one case, it generated a melody and basic chords (hiding, admittedly, somewhere inside a splashy arrangement contributed by a human arranger and performers). In another case, Google AI produced some, shall we say, abstruse piano fragments. But we haven’t seen anything like this before. Singer/songwriter — and YouTube star — Taryn Southern is working on an collection of songs whose instrumental backings were created entirely by an AI program (Southern writes the melodies and lyrics).
The appropriately futuristic, palindromic title for the album is: I Am AI. The first single, “Break Free,” came out a couple of weeks ago.
One of the things that’s immediately striking about “Break Free” is that it sounds so normal. It’s not bloops and bleeps, nor is it... well... as completely insane as previous attempts have seemed.
The AI app Southern is using is called Amper. It’s a web-based tool in which you set some basic parameters for the music you want, and Amper creates, or renders, it for you. If you don’t like what Amper comes up with, you can render again. Its creators promise the app never repeats itself, presenting you with a new chunk of music with each render. Southern tells Big Think that she’s done this as many as 134 times to get what she was looking for in a particular piece of music.
The few criteria you provide Amper are pretty simple. You can choose from among four styles.
You can select your instrumentation.
You can also pick your tempo — in minutes and seconds, though unfortunately not bars — and you can select the key you want.
And that’s it. Amper is not really designed for writing songs, per se. It’s more about generating music backgrounds for videos, games and augmented reality projects. Southern says the company told her she was the first artist using it for songwriting as far as they knew, and they’ve been learning about their own program from what she’s doing with it. One obvious question that comes to mind is who exactly owns the Amper-created songs, Southern or Amper? How about Southern and the person/people who wrote the Amper algorithms? When asked about this, Southern laughs, saying, “We are working through those issues.”
Southern is a self-described “music hacker” who settled on Amper for its high quality after investigating a few other AI-music platforms. She says she loves expressing herself through words and melodies, but after dealing with the expense and logistics involved in recording her music with other humans, she became intrigued with the possibilities of AI: “I got excited that I could do this on my own.”
While the process is cumbersome — especially looking for the right music to fit around a pre-conceived melody, as with “Break Free” — it holds obvious promise. She’s enjoying working with Amper as a creative challenge, thinking of it as a “fun box to play in for an album.” An easier approach would be to let it devise instrumental backings on top of which melodies would be crafted, and that’s a method Southern’s currently exploring. “Right now, I’m writing more to the AI.” She’s also considering layering more human performances on top of the Amper material to add more “interesting textures,” making the AI more “just an inspiration point.”
Amper rendering music (AMPER MUSIC)
There’s a video at the bottom of Amper’s home page that reflects just how new all of this is, and its somewhat hyperbolic claims reveal the unclarity many of us share in trying to envision AI’s rightful place in our lives and in the creative process. Amper’s co-founder and CEO Drew Silverstein says that the creators of Amper believe “the future of music will be created by the collaboration between humans and AI.” It’s hard to know what to make of this.
On one hand, music, art, and other media may be the ultimate forms of human expression, communicating feelings and ideas, as they do, on a unique level. What would AI have to do with that — does AI have a need to express itself?
(ADVENT via SHUTTERSTOCK)
On the other hand, using it as Southern does unquestionably makes sense. You’ll have to decide for yourself, though, if AI is really her collaborator, or simply a tool her own talent and musical taste allow her to effectively employ.
It should also be said that Amper’s primarily aimed at people who express their creativity visually, and simply need original (and free) music that complements their vision, “instantly and with no experience required.” This may be why you set the length of Amper music in minutes and seconds, to match a video’s length. Statements in the video like “We are enabling millions and millions of people to express themselves, and to express their creativity…” and promoting the benefits of “lowering the bar” only really make sense if Silverstein’s addressing visually oriented people.
It would seem that AI music generation would be of the least use to a musician who specializes in composing, and wouldn’t want to shift the responsibility elsewhere. For that person, lyric-writing AI could be a heaven-send. Still, in 2017, it’s as likely to be just about as ridiculous as the movie script we mentioned earlier.
Southern points out an issue that comes from making the creative process so frictionless. “There’s a rub in using something like AI to fuel a creative project,” she says, because, “there’s no need to learn how instrumentation works. We’re seeing this across every creative field. Does it make the person less creative? Does it bring the art form down? I don’t know.” Still, she says, “I’d rather see people have tools to express themselves than not, because that’s an entry point into a different way of thinking and a different way of expressing.”
Ready for the future (TARYN SOUTHERN)
When AI is really fully capable of creating entire songs and recordings — with music, lyrics, and performances — it’s likely we’ll all be at least initially curious to hear what it sounds like. And the field is still so young. As AI becomes more and more alive and individuated — we’re thinking Data on Star Trek TNG — it may well have something very much like feelings and a soul. And it may have something of its own that it absolutely has to express.
"You dream about these kinds of moments when you're a kid," said lead paleontologist David Schmidt.
- The triceratops skull was first discovered in 2019, but was excavated over the summer of 2020.
- It was discovered in the South Dakota Badlands, an area where the Triceratops roamed some 66 million years ago.
- Studying dinosaurs helps scientists better understand the evolution of all life on Earth.
David Schmidt, a geology professor at Westminster College, had just arrived in the South Dakota Badlands in summer 2019 with a group of students for a fossil dig when he received a call from the National Forest Service. A nearby rancher had discovered a strange object poking out of the ground. They wanted Schmidt to take a look.
"One of the very first bones that we saw in the rock was this long cylindrical bone," Schmidt told St. Louis Public Radio. "The first thing that came out of our mouths was, 'That kind of looks like the horn of a triceratops.'"
After authorities gave the go-ahead, Schmidt and a small group of students returned this summer and spent nearly every day of June and July excavating the skull.
Credit: David Schmidt / Westminster College
"We had to be really careful," Schmidt told St. Louis Public Radio. "We couldn't disturb anything at all, because at that point, it was under law enforcement investigation. They were telling us, 'Don't even make footprints,' and I was thinking, 'How are we supposed to do that?'"
Another difficulty was the mammoth size of the skull: about 7 feet long and more than 3,000 pounds. (For context, the largest triceratops skull ever unearthed was about 8.2 feet long.) The skull of Schmidt's dinosaur was likely a Triceratops prorsus, one of two species of triceratops that roamed what's now North America about 66 million years ago.
Credit: David Schmidt / Westminster College
The triceratops was an herbivore, but it was also a favorite meal of the Tyrannosaurus rex. That probably explains why the Dakotas contain many scattered triceratops bone fragments, and, less commonly, complete bones and skulls. In summer 2019, for example, a separate team on a dig in North Dakota made headlines after unearthing a complete triceratops skull that measured five feet in length.
Michael Kjelland, a biology professor who participated in that excavation, said digging up the dinosaur was like completing a "multi-piece, 3-D jigsaw puzzle" that required "engineering that rivaled SpaceX," he jokingly told the New York Times.
Morrison Formation in Colorado
James St. John via Flickr
The Badlands aren't the only spot in North America where paleontologists have found dinosaurs. In the 1870s, Colorado and Wyoming became the first sites of dinosaur discoveries in the U.S., ushering in an era of public fascination with the prehistoric creatures — and a competitive rush to unearth them.
Since, dinosaur bones have been found in 35 states. One of the most fruitful locations for paleontologists has been the Morrison formation, a sequence of Upper Jurassic sedimentary rock that stretches under the Western part of the country. Discovered here were species like Camarasaurus, Diplodocus, Apatosaurus, Stegosaurus, and Allosaurus, to name a few.
|Credit: Nobu Tamura/Wikimedia Commons|
As for "Shady" (the nickname of the South Dakota triceratops), Schmidt and his team have safely transported it to the Westminster campus. They hope to raise funds for restoration, and to return to South Dakota in search of more bones that once belonged to the triceratops.
Studying dinosaurs helps scientists gain a more complete understanding of our evolution, illuminating a through-line that extends from "deep time" to present day. For scientists like Schmidt, there's also the simple joy of coming to face-to-face with a lost world.
"You dream about these kinds of moments when you're a kid," Schmidt told St. Louis Public Radio. "You don't ever think that these things will ever happen."
Are "humanized" pigs the future of medical research?
The U.S. Food and Drug Administration requires all new medicines to be tested in animals before use in people. Pigs make better medical research subjects than mice, because they are closer to humans in size, physiology and genetic makeup.
In recent years, our team at Iowa State University has found a way to make pigs an even closer stand-in for humans. We have successfully transferred components of the human immune system into pigs that lack a functional immune system. This breakthrough has the potential to accelerate medical research in many areas, including virus and vaccine research, as well as cancer and stem cell therapeutics.
Existing biomedical models
Severe Combined Immunodeficiency, or SCID, is a genetic condition that causes impaired development of the immune system. People can develop SCID, as dramatized in the 1976 movie “The Boy in the Plastic Bubble." Other animals can develop SCID, too, including mice.
Researchers in the 1980s recognized that SCID mice could be implanted with human immune cells for further study. Such mice are called “humanized" mice and have been optimized over the past 30 years to study many questions relevant to human health.
Mice are the most commonly used animal in biomedical research, but results from mice often do not translate well to human responses, thanks to differences in metabolism, size and divergent cell functions compared with people.
Nonhuman primates are also used for medical research and are certainly closer stand-ins for humans. But using them for this purpose raises numerous ethical considerations. With these concerns in mind, the National Institutes of Health retired most of its chimpanzees from biomedical research in 2013.
Alternative animal models are in demand.
Swine are a viable option for medical research because of their similarities to humans. And with their widespread commercial use, pigs are met with fewer ethical dilemmas than primates. Upwards of 100 million hogs are slaughtered each year for food in the U.S.
In 2012, groups at Iowa State University and Kansas State University, including Jack Dekkers, an expert in animal breeding and genetics, and Raymond Rowland, a specialist in animal diseases, serendipitously discovered a naturally occurring genetic mutation in pigs that caused SCID. We wondered if we could develop these pigs to create a new biomedical model.
Our group has worked for nearly a decade developing and optimizing SCID pigs for applications in biomedical research. In 2018, we achieved a twofold milestone when working with animal physiologist Jason Ross and his lab. Together we developed a more immunocompromised pig than the original SCID pig – and successfully humanized it, by transferring cultured human immune stem cells into the livers of developing piglets.
During early fetal development, immune cells develop within the liver, providing an opportunity to introduce human cells. We inject human immune stem cells into fetal pig livers using ultrasound imaging as a guide. As the pig fetus develops, the injected human immune stem cells begin to differentiate – or change into other kinds of cells – and spread through the pig's body. Once SCID piglets are born, we can detect human immune cells in their blood, liver, spleen and thymus gland. This humanization is what makes them so valuable for testing new medical treatments.
We have found that human ovarian tumors survive and grow in SCID pigs, giving us an opportunity to study ovarian cancer in a new way. Similarly, because human skin survives on SCID pigs, scientists may be able to develop new treatments for skin burns. Other research possibilities are numerous.
The ultraclean SCID pig biocontainment facility in Ames, Iowa. Adeline Boettcher, CC BY-SA
Pigs in a bubble
Since our pigs lack essential components of their immune system, they are extremely susceptible to infection and require special housing to help reduce exposure to pathogens.
SCID pigs are raised in bubble biocontainment facilities. Positive pressure rooms, which maintain a higher air pressure than the surrounding environment to keep pathogens out, are coupled with highly filtered air and water. All personnel are required to wear full personal protective equipment. We typically have anywhere from two to 15 SCID pigs and breeding animals at a given time. (Our breeding animals do not have SCID, but they are genetic carriers of the mutation, so their offspring may have SCID.)
As with any animal research, ethical considerations are always front and center. All our protocols are approved by Iowa State University's Institutional Animal Care and Use Committee and are in accordance with The National Institutes of Health's Guide for the Care and Use of Laboratory Animals.
Every day, twice a day, our pigs are checked by expert caretakers who monitor their health status and provide engagement. We have veterinarians on call. If any pigs fall ill, and drug or antibiotic intervention does not improve their condition, the animals are humanely euthanized.
Our goal is to continue optimizing our humanized SCID pigs so they can be more readily available for stem cell therapy testing, as well as research in other areas, including cancer. We hope the development of the SCID pig model will pave the way for advancements in therapeutic testing, with the long-term goal of improving human patient outcomes.
Adeline Boettcher earned her research-based Ph.D. working on the SCID project in 2019.
Satellite imagery can help better predict volcanic eruptions by monitoring changes in surface temperature near volcanoes.
- A recent study used data collected by NASA satellites to conduct a statistical analysis of surface temperatures near volcanoes that erupted from 2002 to 2019.
- The results showed that surface temperatures near volcanoes gradually increased in the months and years prior to eruptions.
- The method was able to detect potential eruptions that were not anticipated by other volcano monitoring methods, such as eruptions in Japan in 2014 and Chile in 2015.
How can modern technology help warn us of impending volcanic eruptions?
One promising answer may lie in satellite imagery. In a recent study published in Nature Geoscience, researchers used infrared data collected by NASA satellites to study the conditions near volcanoes in the months and years before they erupted.
The results revealed a pattern: Prior to eruptions, an unusually large amount of heat had been escaping through soil near volcanoes. This diffusion of subterranean heat — which is a byproduct of "large-scale thermal unrest" — could potentially represent a warning sign of future eruptions.
Conceptual model of large-scale thermal unrestCredit: Girona et al.
For the study, the researchers conducted a statistical analysis of changes in surface temperature near volcanoes, using data collected over 16.5 years by NASA's Terra and Aqua satellites. The results showed that eruptions tended to occur around the time when surface temperatures near the volcanoes peaked.
Eruptions were preceded by "subtle but significant long-term (years), large-scale (tens of square kilometres) increases in their radiant heat flux (up to ~1 °C in median radiant temperature)," the researchers wrote. After eruptions, surface temperatures reliably decreased, though the cool-down period took longer for bigger eruptions.
"Volcanoes can experience thermal unrest for several years before eruption," the researchers wrote. "This thermal unrest is dominated by a large-scale phenomenon operating over extensive areas of volcanic edifices, can be an early indicator of volcanic reactivation, can increase prior to different types of eruption and can be tracked through a statistical analysis of little-processed (that is, radiance or radiant temperature) satellite-based remote sensing data with high temporal resolution."
Temporal variations of target volcanoesCredit: Girona et al.
Although using satellites to monitor thermal unrest wouldn't enable scientists to make hyper-specific eruption predictions (like predicting the exact day), it could significantly improve prediction efforts. Seismologists and volcanologists currently use a range of techniques to forecast eruptions, including monitoring for gas emissions, ground deformation, and changes to nearby water channels, to name a few.
Still, none of these techniques have proven completely reliable, both because of the science and the practical barriers (e.g. funding) standing in the way of large-scale monitoring. In 2014, for example, Japan's Mount Ontake suddenly erupted, killing 63 people. It was the nation's deadliest eruption in nearly a century.
In the study, the researchers found that surface temperatures near Mount Ontake had been increasing in the two years prior to the eruption. To date, no other monitoring method has detected "well-defined" warning signs for the 2014 disaster, the researchers noted.
The researchers hope satellite-based infrared monitoring techniques, combined with existing methods, can improve prediction efforts for volcanic eruptions. Volcanic eruptions have killed about 2,000 people since 2000.
"Our findings can open new horizons to better constrain magma–hydrothermal interaction processes, especially when integrated with other datasets, allowing us to explore the thermal budget of volcanoes and anticipate eruptions that are very difficult to forecast through other geophysical/geochemical methods."