• Sunday, 22 December 2024
Why Is TikTok Parent ByteDance Moving Into Biology, Chemistry And Drug Discovery

Why Is TikTok Parent ByteDance Moving Into Biology, Chemistry And Drug Discovery

TikTok’s Chinese parent is beginning to recruit across the U.S. for experts in science and healthcare disciplines far afield from social media. Its motives are unclear.

ByteDance, the Chinese parent company of TikTok, appears to be ramping up work in fields well beyond the bounds of social media: Biology, chemistry, natural sciences and pharmaceuticals. 

The Beijing-based tech giant is recruiting American talent in computational biology, quantum chemistry, molecular dynamics and physics for its “AI for Drug Design” and “AI for Science” teams, according to LinkedIn posts reviewed by Forbes. ByteDance appears to be hiring for at least 17 of these positions across New York, California and Washington state—taking a swing at local rivals like Meta, Google and Amazon, where similar work is already underway. (Others leading these efforts at ByteDance appear to be based in Chicago, Boston and Beijing, according to LinkedIn.) 

The AI for Drug Design team aims to “revolutionize drug discovery,” says one job listing recruiting PhD students to Silicon Valley. “We are dedicated to pushing the boundaries of AI-driven drug design, tackling complex challenges in protein structure prediction, molecular conformation analysis, and computational protein design. By combining our passion for scientific excellence with the transformative power of AI, we aim to accelerate drug discovery and make a meaningful impact on global healthcare.”

 

ByteDance’s AI for Science group, meanwhile, “has been focusing on tackling challenges in natural sciences, including biology, physics, and materials,” say listings for research scientists in computational biology and quantum chemistry in Seattle. And those interning in molecular dynamics in New York will be responsible for pursuing “novel strategies for computer-aided drug discovery” and advancing “free energy methodologies.”

 
 

Staffers across the board will “co-create a future” with ByteDance, the listings say.

 

What, exactly, that future is to be isn’t clear. It is also unclear how drug discovery or development, and other science-focused efforts, fit into ByteDance’s sprawling repertoire of social media sensations like TikTok and its Chinese counterpart Douyin. ByteDance declined to comment on what its objectives are, how large the AI for Drug Design and Science teams are today and who it may have already hired. But a head of AI for drug design and science started at the company this past summer, per LinkedIn—an expert who is also running UCLA’s General Artificial Intelligence Lab—and the team has already released research on protein design and drug design, and an open-source tool combining structural biology with AI.

TikTok did not respond to requests for comment.


How TikTok could help

Forbes found no indication that these ByteDance initiatives are related to TikTok, and TikTok did not respond to a request for comment about whether they are in any way linked.

But experts said data from TikTok and other platforms owned by ByteDance could be highly valuable for these science initiatives. (And though TikTok has emphasized that it operates largely separately from its Chinese parent, Forbes reporting has repeatedly revealed the extent to which the businesses are entangled and data from one company is often accessible by both.)

“They could be doing large-scale hypothesis generation with all this data, and then they could be feeding that data into Chinese pharmaceutical companies or Chinese weapons manufacturers.”

Eric Perakslis, former chief scientist for informatics at the FDA

In the development of new drugs or medical treatments, experts and pharmaceutical companies typically work around what’s called a “target product profile,” a rubric that helps them distill what their optimal drug would look like, the conditions or key problems they’d address with it and who their ideal patients, candidates or markets are. Experts said TikTok—where many people are drawn to health and wellness content, talk openly about medical issues, and regularly seek information on what has (or hasn’t) worked for others—could serve as an invaluable feedback loop and marketing tool for drug developers. Spending data collected from TikTok Shop and in-app purchases could also be useful.

 

"Based on the kinds of videos that [young TikTok users] watch or the content that they consume, it might provide insights as to conditions that they're facing,” said Douglas Schmidt, associate provost for research and co-director of the Data Science Institute at Vanderbilt University, who noted that many young Americans are medicated for ADHD, anxiety and other mental health conditions. “If you could find a way to sell products that would be appealing to them, you can certainly imagine [building] up a robust pharmaceuticals industry or a competitive product, then advertising it on TikTok.”

Eric Perakslis, a former chief scientist for the FDA who previously helped lead pharmaceuticals R&D for Johnson & Johnson, called this “surveillance capitalism basics: everything that you can extract from those things, they do.”

“Given that they're a social media company, and given that they don't have life science divisions that are consenting patients, taking samples or running mice and rat studies and stuff like that—their own data—I think it's a fair assumption that somehow they're accessing data” from elsewhere, said Perakslis.

This growing arm of ByteDance could be a “very big, large antenna,” he added—a way of pulling in a “big noisy mess of data,” filtering it through experts and then figuring out how to monetize it.

"They could be doing large-scale hypothesis generation with all this data, and then they could be feeding that data into Chinese pharmaceutical companies or Chinese weapons manufacturers or Chinese whatever,” Perakslis said. (ByteDance’s head of research appears to be based in Beijing, per LinkedIn.) Beyond drug discovery or repurposing, he noted the language used in these roles could also portend work in food science, agrochemical, petroleum, fuels science or even bioweapons. “The same technologies that allow us to develop new medicines can be used for bioweapons,” Perakslis said. “It's a very similar set of capabilities.”

 

Finally, experts said the roles signal that TikTok’s owner is looking to channel its AI prowess—embodied best by the unmatched TikTok algorithm that steers its ultra-personalized “For You” feed—into other revenue streams. That hypothesis jibes with the job descriptions: Those joining ByteDance’s “multidisciplinary drug discovery team” will be tasked with “applying inventive algorithms” to its work and developing algorithms for antibody design. Other hires will build “cutting-edge machine learning technologies in scientific areas like biology, physics and materials.”


State of play

This is not the first time that ByteDance—which established the ByteDance AI Lab in 2016—has ventured into industries that seemingly have little to do with TikTok, Douyin and social media more broadly.

Healthcare is just one of many. Last year, ByteDance bought Chinese hospital chain Amcare for $1.5 billion—a push to establish a foothold in the space after ByteDance had acquired online medical encyclopedia Baikemy in 2020 and subsequently launched a suite of healthcare apps under the name Xiaohe. Experts at the time told Forbes those moves reflected the Chinese government’s mandate for the country to be a global leader in health by the end of this decade.

What’s different now, though, is that ByteDance appears to be expanding those efforts beyond China into the United States—hiring scientists and PhDs in at least three major U.S. cities. And in doing so, it’s planting a flag in a market where some of its biggest American competitors have their own science, research or health-focused initiatives underway and edging into their AI research talent pool.

“They’ve attracted a lot of negative publicity about TikTok. … Maybe they’re trying to diversify into fields that are viewed more positively by governments and regulators.”

Douglas Schmidt, co-director of the Data Science Institute at Vanderbilt

Like ByteDance, these U.S. rivals are tech behemoths that are trying to win in industries well outside their bread-and-butter businesses. Google parent Alphabet is pursuing breakthroughs in tech and science through its Moonshot Factory (previously called Google X). Google DeepMind, the search giant’s innovation lab for AI and neuroscience, recently released a model that can mimic human reasoning. Meta AI, a research hub from the parent company of Facebook and Instagram, is transforming medicine by deploying AI to improve and expand access to MRIs and help doctors read and make predictions off X-rays. And Amazon has had its own foray into healthcare, acquiring OneMedical in 2022. (TikTok Shop, which the platform launched in the U.S. in September, is also a growing threat for Amazon.)

“They're undoubtedly trying to find a way to leapfrog or get access to people with skills in those areas,” said Schmidt, the professor from Vanderbilt.


Got a tip about ByteDance or TikTok? Reach out securely to Alexandra S. Levine on Signal/WhatsApp at (310) 526–1242, or email her at alevine@forbes.com.


But ByteDance’s segue into the sciences could also simply be a way for the Chinese company to improve its goodwill in the U.S. as talks of a TikTok ban loom and the app faces heavy scrutiny for its effect on the American public and discourse, he added.

“They've attracted a lot of negative publicity about TikTok and its corruption on the youth of America,” Schmidt said. “But when we start thinking about healthcare and science, that comes across as more corporately responsible and trying to help for the good of humanity. … Maybe they're trying to diversify into fields that are viewed more positively by governments and regulators.”

After all, as ByteDance says itself: “Our goal is to create breakthroughs in natural science with new methodology and help the world.”

Share on

SHARE YOUR COMMENT

// //