It’s certainly possible enough to identify the most obvious astroturf and spam for further investigation – there are plenty of commercial and FOSS analysis tools that do just that. Spambots-for-hire aren’t exactly subtle in how they work, and there are signatures like IP address, country of origin, frequency, spacing, content, time of submission, etc. that in combination can flag a submission (or, more likely in this case, a group of submissions) as highly probable spam for further investigation. One can do that and still err on the side of caution in terms of allowing individual false-positive outliers that may seem like spam but are in fact legit while clearing out the junk.
No, but as part of their jobs they should be able to determine what is essentially a flood of comment spam and advise appointed* officials about the problem and impact on the servers and their mission. You might as well ask “should @orenwolf at BoingBoing be able to determine what constitutes a ‘valid’ submission?” – clearing out comment spam and advising non-tech management about flare-ups are depressingly regular parts of a Web sysop’s job description.
[* FCC heads like Pai are not elected]
