In 2003, David Baker, a biochemist at the University of Washington, designed the first novel protein and predicted its 3-dimensional structure. Although technology and science have long been intertwined, this new connection unintentionally started the scientific community down a long path leading straight into the hands of Big Tech.
Baker, along with his team, did this using a software of his own design called Rosetta. Rosetta uses algorithms to allow scientists to model and analyze the structure and folding of proteins and visualize the way they interact with other molecules. Most notably, this enables “de novo protein design”, or the generation of novel proteins that do not exist in nature.
In recent years, computational chemistry has become more prevalent at Whitman, in part because of Associate Professor of Chemistry Dalia Biswas. Though she has experience in several different fields within chemistry and teaches a wide range of courses, Biswas has been active in bringing computational chemistry to Whitman, for instance designing the Wilke Family Computational Lab that is housed in the Hall of Science. Last year, she became a Data Chemist Network affiliate with the National Science Foundation’s Center for Computer-Assisted Synthesis. She said that the “big idea” of computational chemistry is to increase efficiency and find ways to save labor, time and energy.
Baker’s work has long been admired within the scientific community. Biswas spoke highly of his research, vocalizing her excitement about Rosetta’s impact on the scientific world. Baker’s dedication has earned him many awards commemorating important contributions to the studies of protein folding and protein design. This recognition came to a head on Oct. 9, 2024, when he was awarded the Nobel Prize in Chemistry “for computational protein design”.
Unfortunately, his achievement was quickly overshadowed. Baker was awarded only half of the Nobel Prize, with the other half going to two executives at Google DeepMind, Google’s artificial intelligence research lab.
The awarding of the Nobel Prize in Chemistry to the co-founder and CEO of Google DeepMind Demis Hassabis and John Jumper, the director of Google DeepMind, has stirred up great controversy within the scientific community. It sparks questions about who deserves recognition when an advancement is built off of the work of others, if the promise of “efficiency” is worth the hidden drawbacks and the role of large corporations in science. Explorations of these questions are, much like a protein, multifold and complex.
The technology that Hassabis and Jumper were recognized for is called AlphaFold, an artificial intelligence program that uses a large language model to predict the folding and 3-dimensional structure of proteins. If the ultimate goal of computational chemistry is to maximize efficiency, then AlphaFold may be the greatest thing to ever happen to the field. However, when the curtain is peeled back, it is no longer the holy grail of protein discovery that it seems to be.
Another prominent figure in the sciences at Whitman weighed into the debate with fascinating insight. Associate Professor of Chemistry Tim Machonkin knows firsthand the blood, sweat and tears that go into being a part of a chemistry research lab, having spent about 35 years working in the field of bioinorganic chemistry. His time in research labs at institutions such as Stanford and the University of Wisconsin, as well as his current research which focuses on metal-containing enzymes, informs Machonkin’s perspective on the matter at hand.
Machonkin argues that, although useful, the work done by Hassabis and Jumper is not novel.
“AlphaFold does not, from what I can tell, represent an advance in the fundamental theory of protein folding and protein structure,” Machonkin said.
He places this in contrast to work done by other scientists, such as David Baker, putting forth a sentiment that Biswas agrees with as well.
“David Baker’s award is a different story—his is definitely well-earned, and not a surprise,” Machonkin said.
Conversely, Hassabis and Jumper were able to achieve their feat because of Google. Though AlphaFold can make predictions more accurately than earlier programs such as Rosetta, it is only able to do so because of the amount of work that scientists have put into the algorithms and models that preceded it. The entire function of large language models is based on the ability to mine and sort through pre-existing data. In other words, it could only be put together by a partnership like Hassabis and Jumper, who are privy to the resources and funding of a company like Google– a company that has such a chokehold on the internet and online search that it was ruled “monopolist” in an antitrust lawsuit filed by the U.S. Department of Justice in August of this year.
This case is only one of the many ongoing antitrust lawsuits against Google filed by the Department of Justice and multiple states, who sued Google over accusations that they have illegally maintained a monopoly as a search engine. They have done so by paying companies such as Apple and Mozilla, to name a few, in order for Google to remain the default search engine on these devices– in a single year, 2021, Google’s search engine dominance expenditures reached a whopping $26.3 billion. Thanks to these payouts, Google enjoys a nearly 90 percent share of the overall search engine market, a number that increases to approximately 95 percent when looking at only mobile devices.
After it was ruled that Google violated the Sherman Antitrust Act, which protects free competition and prevents unlawful monopolies, they immediately went on the defensive. Alphabet, Google’s parent company, released a statement asserting that the reason that they dominate the market is because of the quality of their services, not because a portion of their $305.6 billion annual revenue helps other wealthy corporations– corporations who are also currently being sued by the Department of Justice for violating antitrust laws. In their statement, Alphabet goes on to suggest that people’s access to online search would be hindered were these unlawful payouts to stop.
“This decision recognizes that Google offers the best search engine, but concludes that we shouldn’t be allowed to make it easily available,” Alphabet said in a statement they released shortly after their case verdict.
In conversations with the Whitman Chemistry Department, Biswas and Machonkin both touched on the unfairness that this year’s Nobel Prize represents. Machonkin is unsurprised that AlphaFold emerged from a company such as Google.
“At Google, Hassabis and Jumper had access to essentially unlimited funds and unlimited computational resources,” Machonkin said.
Many scientists like Baker who have investigated protein structure throughout the years could have, in theory, been able to put a system like AlphaFold together. The vast majority of researchers have to work from the ground up, spending time fighting to get grants and any other funding they can get access to. They, like Baker, have the misfortune of not being employed and backed by a corporation worth multiple trillions of dollars and with databases filled with findings and work done by scientists who deserve this support and recognition.
And yes, awards such as the Nobel Prize have always been “unfair” to some extent. Out of the 192 people awarded the Nobel Prize in Chemistry, 151 of them have hailed from the United States, the United Kingdom and Germany. These are also the three countries who spend the most money on this type of research. 2024 also is not the first year that the recognition has gone to research done in a corporate lab rather than the academic research that is often seen as the standard– the 1956 Nobel Prize in Physics was awarded for work on the transistor, which was done in AT&T’s Bell Labs. However, the 2024 Nobel Prize in Chemistry as well as this year’s Nobel Prize in Physics, which went to Geoffrey Hinton “for foundational discoveries and inventions that enable machine learning with artificial neural networks”, set a precedent for Big Tech’s tightening grip on the scientific world. While its looming presence does allow for increasing efficiency of scientific discovery, there are also great drawbacks to consider.
These scientific advancements come at the cost of allowing companies to operate in ways that contradict what is lawful in the eyes of the federal government. These corporations dominate our searches, our information, our advertisements and our attention. Are the medical and scientific benefits that will come from programs such as AlphaFold worth the devaluation of the researchers who have dedicated lifetimes to the findings that are foundational to these technologies? Even experts within the community are torn on the issue.
After considering the advantages and disadvantages of the technology itself, as well as the impact of Google’s role in the discovery, Machonkin put his thoughts on AlphaFold’s impact on the future nicely.
“What does this mean for the future? Well, I think a program like AlphaFold represents one of the extremely few legitimate and ethical uses for AI. The other 99.9 percent is total garbage, and exceedingly dangerous garbage at that. Just like how today we scratch our heads wondering, ‘how did people in the 1950s think mass spraying of DDT was a good idea?’ the historians 75 years from now, if there are any, will be scratching their heads wondering, ‘how did people in the 2020s think mass release of AI was a good idea?’” Machonkin said.
