AMI used for "aws batch compute environment" when "image_id" is not defined

AWS Linux AMI standard support ended last Decemeber.

The Amazon Linux AMI has ended standard support as of December 31, 2020. … Customers are encouraged to upgrade their applications to use Amazon Linux 2, which includes long term support through 2023.

I have an existing service using aws_batch_compute_environment where the (optional) image_id parameter is not defined.

Is it safe to assume that the underlying AMI will be auto-updated to use the newer AWS Linux 2? Or do I need to update my environment to specify the image_id in order to assure that my service is using a supported AMI?

(I didn’t see in the terraform documentation how this is handled)