TikTok CEO on AI: Breakthroughs in Three Years and the Risks He Named (2026)

The TikTok CEO on AI says science is about to feel it. Speaking in Singapore on October 12, 2026, Shou Zi Chew said he expects artificial intelligence to drive real scientific breakthroughs within two to three years, and he also named the cybersecurity and agent risks that worry him. This guide to the TikTok CEO on AI sets out what he said, where he said it and what the reporting does not cover.

Quotes below come from CNBC’s report of the talk, published the same day, and the event details from DBS’s announcement of the conference. What the TikTok CEO on AI said is a CEO’s opinion on a stage, not a TikTok product announcement.

What to know:

• The prediction: “In the next two to three years, we’re going to have a real scientific breakthrough: physics, biotech, chemistry, material science, because of AI,” Chew said, per CNBC.
• The caution: he called the cybersecurity risks “quite big” and warned that users may give AI agents extensive permissions without understanding the consequences.
• The setting: the DBS Asia Future Forward Conference in Singapore, a fireside discussion on unlocking human potential in an AI era.
• Not covered: CNBC’s report contains no new TikTok product, policy or date.

What the TikTok CEO Said About AI

Chew said he sees AI’s progress in science “trending on a very steep curve.” In his words: “The trend is very very clear. It’s that we are only at the beginning of what artificial intelligence can probably do.”

He separated the short term from the long term. “I think in the short term we are probably overestimating it, but in the long term we are probably underestimating it,” he said, per CNBC. CNBC also reports he said rapid advances were opening up possibilities far beyond the technology’s current applications, “even as questions remain over how to ensure increasingly powerful systems operate within human ethical boundaries.”

On that last point he asked: “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?”

The Risks the TikTok CEO on AI Named

Chew said AI’s growing capabilities will require stronger safeguards, particularly as autonomous systems spread. “I think the cybersecurity risks are quite big,” he said.

• Permissions: users may grant AI agents extensive permissions without fully understanding the potential consequences, per CNBC.
• Purchases: he pointed to agents that buy things autonomously, saying such systems could end up buying more than users intended in their eagerness to complete a task.
• Fraud costs: as AI models get more sophisticated, they make it cheaper for malicious actors to carry out fraud. “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 said.

Where the TikTok CEO on AI Spoke

DBS, the Singapore bank, announced the DBS Asia Future Forward Conference as a new flagship event held in Singapore on October 12, 2026, with AI at the center of the agenda and some 700 DBS corporate and wealth clients expected. Its announcement lists Chew for a fireside discussion with a DBS executive on unlocking human potential in an AI era.

The conference’s stated theme was “intentional AI,” which DBS defines as deploying AI with a clear and specific purpose and ensuring the technology creates useful, measurable impact. Chew’s remarks as reported fit that frame: a bet on science paired with a list of things that could go wrong.

ByteDance’s Science Push

CNBC ties the remarks to ByteDance’s work beyond consumer apps. It reports that 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. Chew said there is “an effort to see how we can collaborate with a lot of scientists to make some progress in science as well.”

CNBC adds that AI is reshaping TikTok’s operations “from content creation and recommendations to search and moderation,” and that ByteDance is developing its own AI models.

CNBC places the remarks in a wider moment. It notes that last week the US 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. That initiative is not a TikTok project, and CNBC cites it as context for the industry’s spending on AI-powered science.

How It Relates to TikTok’s Own AI Shopping Tools

Chew’s example of agents buying more than intended is a general warning about the technology. CNBC does not report that he referred to any TikTok product. For context, TikTok announced a conversational Shopping Assistant on October 5 that it says can handle shipping, sizing, availability and purchasing in one conversation, as covered in TikTok Buy Direct and Shopping Assistant Explained. That announcement gave no rollout date.

What Agentic AI Means for the TikTok CEO on AI

“Agentic” is the word that sits behind Chew’s warning about permissions and purchases. It means AI that takes actions for a user or a business, not just answers questions. An agent that can act needs access to accounts, payment details or tools, which is why the permission question he raised is the practical one: what the system can touch matters as much as what it can say.

CNBC’s report quotes Chew describing the specific worry of autonomous buying: the systems “could end up buying more than users intended” in their eagerness to complete a task, as CNBC puts it. That is a statement about how agents are built to finish what they start, and it is the reason he ties the topic to safeguards.

TikTok’s Own AI Tools This Month

The TikTok CEO on AI spoke in general terms, but TikTok has been shipping agent-style tools of its own. In its October 5 Newsroom post, TikTok announced three items that fit the “agentic” label, all covered in this site’s Buy Direct and Shopping Assistant guide.

• Shopping Assistant: a conversational AI agent that TikTok says handles product details, shipping, sizing, availability and purchasing in one conversation. TikTok gave no rollout date.
• Agentic Leads: an AI agent for lead generation that holds guided conversations to “capture, qualify and convert high-intent, actionable consumer prospects,” planned for TikTok Direct Message and advertisers’ lead forms.
• TikTok for Business MCP: an interface that lets AI tools manage ad campaigns. TikTok reports an “over 200% increase in advertisers using our MCP” and says it is now available as a connector across Claude, Kimi, Manus, Perplexity, Replit, Snowflake, Tencent WorkBuddy and the IAB Tech Lab Agent Registry.

None of this appears in CNBC’s report of the Singapore talk, and this guide does not claim Chew was speaking about these products. The point is simply that the permissions and purchasing questions he raised are live product questions at TikTok this month.

Reading the TikTok CEO on AI: the Quotes

Two lines in the talk carry most of the weight. The first is the near-term claim: “In the next two to three years, we’re going to have a real scientific breakthrough.” It is a timeline and a field list, physics, biotech, chemistry and material science, stated as “very, very, very possible” rather than certain.

The second is the balancing line, overestimated in the short term and underestimated in the long term. Read together with his cybersecurity comments, it frames AI as something that deserves both patience and guardrails. Both lines are Chew’s opinion, and CNBC’s report gives no data behind the timeline.

This site’s suggestion, not Chew’s, is a smaller and more practical takeaway for TikTok users. If an AI feature asks for permissions, such as access to payment methods or the ability to act on your behalf, it is worth reading what it can do first. For protecting your own TikTok account, TikTok’s own steps are in this site’s Account Security guide.

What Is Not Known

• The full talk: this guide to the TikTok CEO on AI relies on CNBC’s report, so anything Chew said that CNBC did not report is not covered.
• Product plans: the report mentions no TikTok feature, launch or policy tied to these remarks.
• Timeline basis: the “two to three years” estimate is Chew’s own view, and the report gives no research behind it.

Frequently Asked Questions

What did the TikTok CEO on AI say at the DBS conference?

Shou Zi Chew said that in the next two to three years AI will produce real scientific breakthroughs in areas such as physics, biotech, chemistry and material science, and that cybersecurity risks are “quite big.”

Where did Shou Chew say this?

At the DBS Asia Future Forward Conference in Singapore on October 12, 2026, per CNBC.

What risks did the TikTok CEO name?

Users granting AI agents extensive permissions without understanding the consequences, agents buying more than users intended, and falling costs of fraud for bad actors.

Is TikTok launching an AI product?

The CNBC report on the remarks does not mention a new TikTok product. TikTok’s October 5 Shopping Assistant announcement is a separate item.

Sources: CNBC, October 12, 2026, DBS announcement via Thailand Business News.