Except that this is very, very normal in almost all research that requires funding. Talk to faculty members in engineering departments, and they'll tell you they often spend most of their professional time grant writing.

Saying this has a chilling effect is like saying a store requiring money in exchange for goods has a chilling effect because it means the purchaser is constantly in need of earning money to eat.

If they give someone money to support research without specific goals, playing nice is implied.

Unprofessional conduct (blowing money on donuts) is reasonably expected to not attract more funds.

But when does undesirable conduct come into play? Would they fund a guy who discovered that YouTube improves high school academic performance by 48%? Would they fund a guy who finds that YouTube slows child cognitive development?

You don't know, because the company's decision making is unknown. But a clever professor who needs grants could probably make a good guess. Even if Google (or any other company) doesn't act that way or has no desire to influence research, the perception of bias still exists, and that undermines science.

My local library is prohibited from accepting many kinds of gifts as a government entity. So the library board setup a "Friends of the Library" non-profit with its own board that can accept and distribute cash or other gifts, subject to reporting requirements. You need a similar concept where you give up control.

They key thing here is that perception is more important than reality. You need to work to build and maintain trust.