Concern about AI risks is growing.

A global answer needs the U.S. and China to work together.

But the two big powers trust each other less and less.

They do not trust each other's AI plans.

AI rules may come up at a meeting next week.

President Donald Trump and Chinese leader Xi Jinping will meet in Washington.

Both countries want to win at AI.

Trump warned that rules on the technology would help China.

He said the leading nation "wins."

Samm Sacks works at Johns Hopkins.

She said a formal deal may be almost impossible.

But Trump and Xi only need to say AI is risky for both countries.

The two countries grow AI in different ways.

U.S. companies build frontier models.

They use top chips and more data centers.

China pushes its open-source models around the world.

These models are cheaper.

They are growing with U.S. companies even with U.S. chip limits.

Treasury Secretary Scott Bessent spoke on Tuesday.

He said the U.S. needs more open-source models to answer China.

The U.S. says China tries to take skills from U.S. AI systems.

China says the U.S. wants an "AI monopoly."

Trump adviser David Sacks said China has "already rejected" slowing AI work.

He said promises cannot be checked.

Still, former Chinese ambassador Cui Tiankai spoke at a Beijing conference.

He said dialogue is needed.

Anthropic and OpenAI both limit access in China.

They cite national security.

But they also call for teamwork on safety.