Removal of embedded tools can cause harms through: (1) failure to address changes that have been made that rely on the tool; and (2) failure to adequately address the ongoing impacts a tool may have on wider systems.'Removal' may mean actually taking a tool out of a context, but it could also mean no longer supporting an existing tool, increasing resource needs resulting in a tool becoming unaffordable (financially or otherwise), incompatabilities with new tools, etc.An example of the first kind of harm is e.g. a company no longer supporting a medically implanted device for vision loss patients https://spectrum.ieee.org/bionic-eye-obsoleteAn example of the second kind of harm is the ongoing impacts from the UK Exam moderation algorithm in 2020, which had knock-on impacts internationally even after the algorithm itself was removed https://dl.acm.org/doi/abs/<br
1. 1145/
1. 3533186>Some of these concerns, and possible approaches to them, are discussed in this (open access) article: Gold, N. E., Lawson, I., & Oxtoby, N. P. (2023). Afterlife: the post-research affect and effect of software. Research Ethics, 0(0). https://doi.org/<br
1. 1177/17470161231178450>