TikTok CEO Shou Zi Chew expects artificial intelligence to produce genuine scientific breakthroughs in physics, biotech, chemistry and materials science within the next two to three years — a timeline he described as not just possible, but likely.
Chew made the prediction at the DBS Asia Future Forward Conference in Singapore on Monday, according to CNBC, offering one of the most concrete timelines yet from a major technology executive on when AI will move from powering apps to accelerating discovery in the hard sciences. For readers tracking how quickly the field is moving, the latest AI news increasingly reads like a ledger of claims that would have sounded fanciful a year ago.
"In the next two to three years, we're going to have a real scientific breakthrough: physics, biotech, chemistry, material science, because of AI. It's very, very, very possible," Chew said.
ByteDance Moves Beyond Consumer Apps
Chew's remarks come as ByteDance, TikTok's Chinese parent company, expands its artificial intelligence research beyond consumer applications, CNBC reported. In April, the company's Seed research team unveiled Protenix-v2, an AI model focused on biomolecular structure prediction and design, with potential applications in drug discovery.
The executive signaled that more scientific work is on the way. "There's an effort to see how we can collaborate with a lot of scientists to make some progress in science as well," Chew said. "There's been some work that we've done in this area, and it is very exciting."
That scientific ambition coexists with a transformation already underway inside TikTok itself. Chew noted that AI increasingly reshapes the platform's operations across content creation, recommendations, search and moderation — the core machinery of a service used by hundreds of millions of people daily.
Only the Beginning of a Very Steep Curve
Chew framed the current moment as an inflection point rather than a peak, arguing that the pace of progress — particularly in scientific applications — is accelerating faster than most observers appreciate.
"The trend is very very clear. It's that we are only at the beginning of what artificial intelligence can probably do," he said. "And if you look at where the trend is, [where] the progress is, particularly in areas of science, it's really trending on a very steep curve at this moment in time."
He also echoed a formulation that has become common among AI leaders: near-term hype exceeds reality, while long-term impact is systematically underestimated. "I think in the short term we are probably overestimating it, but in the long term we are probably underestimating it," Chew said.
Big Tech Is Already Betting on AI-Driven Biology
Chew's timeline lands amid tangible movement from governments and rival platforms. Last week, the U.S. government, Meta and Google announced their participation in a $1.8 billion initiative to develop AI models capable of predicting how cells behave, in an effort to accelerate biological research and drug discovery, according to CNBC.
Against that backdrop, ByteDance's interest in computational biology — of which Protenix-v2 is the clearest public evidence — looks less like a side project and more like a bid to compete where AI's next measurable wins are expected to land. If Chew is right about the two-to-three-year window, the companies positioned to claim those wins will be those already collaborating with working scientists, not just those training larger general-purpose models.
The Alignment Question Chew Can't Shake
For all his optimism, Chew acknowledged that powerful scientific AI raises a problem the industry has not solved: making sure increasingly capable systems stay within human intentions.
"If that happens, then you know the question would be: how do you align the models? How do you align the machines to be doing things that are obvious to us as human beings?" he said.
It is a notable admission from an executive whose company is racing to expand what its models can do. The alignment challenge Chew described is generally understood to grow harder as systems gain autonomy — and TikTok's parent is pushing exactly in that direction.
A Platform CEO's Cybersecurity Warning
The most cautionary portion of Chew's remarks concerned AI agents — systems that act on a user's behalf — and the security exposure they create for ordinary users and platforms alike.
"I think the cybersecurity risks are quite big," Chew said, warning that users may grant AI agents extensive permissions without fully understanding the potential consequences. He pointed to the risks of letting agents make purchases autonomously, noting that the systems could end up buying more than users intended in their eagerness to complete a task.
Chew also flagged what falling costs mean for attackers: increasingly sophisticated AI models make it significantly cheaper for malicious actors to carry out fraudulent activities. "So you need the right tools to now you know defend against the fast falling cost of creating problems for platforms like us, security problems for us," he emphasized.
The warning carries weight coming from a platform that sits in the crosshairs of large-scale abuse attempts. If the cost of producing convincing fraud, bots and manipulation continues to fall, platform defenses will need to scale at least as fast — a proposition Chew seemed doubtful about leaving to chance.
What to Watch
Chew's prediction is testable, which is what makes it notable. Within roughly three years, either AI-assisted research produces a breakthrough of genuine significance in physics, biotech, chemistry or materials science — or one of the industry's most prominent executives will have oversold the timeline, as tech leaders have sometimes done before.
Either outcome will shape how governments, investors and scientists calibrate their expectations. The immediate question for ByteDance is whether its scientific collaborations move from aspiration to published results. The broader question is whether the rest of the industry's bullish timelines start converging on Chew's — and whether the safeguards he says are needed arrive before the agents do.
---
Stay Ahead of AIGet the latest AI news, analysis, and breakthroughs — all in one place.
Read more AI news →
