Checkout charity by default in an online cart

John is responsible for designing an online shopping cart and checkout experience at a mid-size organization.

John goes to the grocery store, where a checkout clerk asks if he'd like to donate a few dollars to a good cause. Let's suppose, for the sake of argument, that John and his community recognize it as a good cause.

John goes back to work, does some research, and finds that these "checkout charity" asks bring in hundreds of millions of dollars annually in just his country, which is the US.

John has an idea:
"Let's add this feature to our online shopping cart! On the checkout page, we'll ask folks to donate a few dollars to a good cause." Knowing the power of defaults, he even decides "Let's make that the default, for every checkout. We'll also provide a link someone can click on to get a popup explaining what the donation is for, and allowing them to change the amount, increasing it up quite a bit or decreasing it, even down to zero if they want. Because they have the option of setting the donation to zero, it's completely optional, but because we're making it easier to give, and a lot of the people who use our checkout are likely to be of the giving type and see this as a good cause, we'll be able to collect a lot more donations this way, compared to if we made the default zero. We'll give folks a receipt they can use for a US tax deduction. Since it's online and we provide more information about the charity than at the grocery store, it doesn't even have the 'public shaming' or non-transparency issues that some cheapskates grouse about online, plus we have customers' information to enable re-contact, and we'll pick a charity that aligns with what people are going through the checkout for. Even if some Canadian public radio types don't like this, at least some polls suggest most Americans do. "

Suppose that once this change is made, donations per checkout do go up, quite significantly.

Are there any significant drawbacks John should be aware of? If so, is there any evidence to suggest whether they are or aren't strong enough to outweigh the benefits? Are there other benefits? Are there studies that have been done about this, looking at long-term impacts or other effects?