I do not get the negativity right here.
Within the paper they suggest the long run aim:
SENSING ALCOHOL INTOXICATION in actual time
might provide alternatives for triggering just-in-time interventions
geared toward bettering prevention and remedy of
alcohol use issues. For people in remedy, it might
set off speedy distant help from a sponsor, which
might scale back relapse threat. For a person with heavy
ingesting, it might set off just-in-time assets geared toward
decreasing additional alcohol consumption or different associated dangers
like ingesting and driving, probably mitigating particular person
and public well being harms.
So an app which could cease somebody from drunk driving or ingesting to a lot.
To create such an app you begin growth clearly with optimum knowledge in good circumstances after which generalize. So that you first get good knowledge by strapping the telephone to their again. In a while you may put it of their pocket and attempt to get knowledge of comparable high quality by filtering the noisy knowledge launched by the telephone shifting and shaking round.
With the recorded optimum knowledge you strive a number of issues and spot:
“Hey it really works with personalised gait profiles!” which is completely nice as a result of often one individual makes use of one telephone. Because you’re in a analysis division who usually wrestle to maintain the cash flowing, you clearly publish your outcomes (that are straightforward to market/clickbait)
Getting your analysis into the information is a good way to get extra money, which you’ll then use to:
Which leads you nearer to the unique aim.
That is how science will get accomplished, and sadly has to get accomplished, cope with it.