We can (at least for the purpose of argument) take ‘unethical’ to mean ‘deceptive’, in which case ‘is automatic authorship unethical?’ can be read as ‘is automatic authorship deceptive?’ But that depends on who, in the readership, is being deceived, and different disciplines may ‘read’ the author list differently.

My presence in an author list may indicate a claim that:

1. I contributed to this paper, and my position in the list order reflects the extent of my contribution.
2. I contributed to this paper, and my position in the list reflects the alphabetisation of my name.
3. I at least read this paper, and my inclusion in the list reflects the fact that I, along with 800 other people, was during the relevant period funded 50% or more of my time to work on this collaboration (actually, in this case you can't really tell whether or not I'm even aware the paper exists).
4. I'm aware of this paper (which happened in my lab) and am the PI who got the funding (I get the impression, as others have remarked, that this is common in biomedical fields).

As long as the readership of the paper parses the author list in the way I expect, then no-one has been deceived, even if someone from a different area, who parses author lists differently, would get the wrong impression, and might think I was claiming something I'm not (and therefore reproach me for being unethical in doing so).

So – I think the answer to the question ‘is “automatic authorship” unethical?’ is: it depends on what you mean by ‘automatic’, ‘authorship’ and ‘unethical’ (we can perhaps take ‘is’ as unproblematic), and depends crucially on what author and reader believe is being communicated by one's presence and location in an author list.

I've written papers under each set of rules apart from the fourth. In one similar case, however (anecdote alert), I was on a five-author paper where almost all of the actual writing was done by the PI of the grant and me, in that order; the PI sequenced me first in the author list, then the second, third and fourth authors, then him last, to log the fact that he was the PI, even though you couldn't then tell how much he'd written. I didn't think this was very fair to him, but he claimed this was the norm in his discipline (a different area of physics).

Some disciplines seem to observe more than one of these parsing rules.

Of course, all of this means that bibliometrics which count things like ‘first-author papers’ are so errorprone as to be essentially pointless, for this as for so many other reasons.

I don't think one can claim this is ‘unethical’ in general.

We can (at least for the purpose of argument) take ‘unethical’ to mean ‘deceptive’, in which case ‘is automatic authorship unethical?’ can be read as ‘is automatic authorship deceptive?’ But that depends on who, in the readership, is being deceived, and different disciplines may ‘read’ the author list differently.

My presence in an author list may indicate a claim that:

1. I contributed to this paper, and my position in the list order reflects the importance of my contribution.
2. I contributed to this paper, and my position in the list reflects the alphabetisation of my name.
3. I contributed to this paper, and my position in the list represents the order in which the various authors first committed edits to the text in github.
4. I at least read this paper, and my inclusion in the list reflects the fact that I, along with 800 other people, was during the relevant period funded 50% or more of my time to work on this collaboration (actually, in this case you can't really tell whether or not I'm even aware the paper exists).
5. I'm aware of this paper (which happened in my lab) and am the PI who got the funding (I get the impression, as others have remarked, that this is common in biomedical fields).

As long as the readership of the paper parses the author list in the way I expect, then no-one has been deceived, even if someone from a different area, who parses author lists differently, would get the wrong impression, and might think I was claiming something I'm not (and therefore reproach me for being unethical in doing so).

So – I think the answer to the question ‘is “automatic authorship” unethical?’ is: it depends on what you mean by ‘automatic’, ‘authorship’ and ‘unethical’ (we can perhaps take ‘is’ as unproblematic), and depends crucially on what author and reader believe is being communicated by one's presence and location in an author list.

I've written papers under each set of rules apart from the fifth. In one similar case, however (anecdote alert), I was on a five-author paper where almost all of the actual writing was done by the PI of the grant and me, in that order; the PI sequenced me first in the author list, then the second, third and fourth authors, then him last, to log the fact that he was the PI, even though you couldn't then tell how much he'd written. I didn't think this was very fair to him, but he asserted this was the norm in his discipline (a different area of physics).

Some disciplines seem to observe more than one of these parsing rules.

Of course, all of this means that bibliometrics which count things like ‘first-author papers’ are so errorprone as to be essentially pointless, for this as for so many other reasons.