AI helps fight the replication crisis in science

That is exactly the kind of protocol I had follow for 

years when I was doing my dissertation research --

it was very tedious, requiring constant focus on every little thing

As we all know, the sciences, especially the social and biomedical sciences, are fraught with published research that cannot be replicated and is thus either questionable or debunked. This is a huge problem. At least scientists no longer dismiss the problem as minor. It is major. Steps have been taken to reduce non-replicability, e.g., by posting detailed research and data protocols online before conducing the research. That has reduced the plague of p-hacking[1] to what appear to be low levels. Other changes include now requiring scientists to post all of their data and detailed protocols online so that other labs can re-analyze the data to see if there are mistakes in the data analysis. At least scientists are giving it the good old college try!  

🤪

A SciAm article, AI-powered smart goggles are helping novice scientists perform like experts, describes a new way to attack the always present problem of human error in conducting research protocols. It is easy to make a mistake when you are doing a complex, many-step research protocol. One tiny error, e.g., skipped a step, or added too much or too little reagent to a mixture, etc., can and usually does completely wreck the outcome of tedious work that took a lot of time and money to generate.

Short story made shorter, researchers wear AI goggles that have the research protocol fed into them. The AI then "watches"  the researcher conduct all of the steps in the protocol. If a mistake is about to be made or has been made, the AI will tell the researcher they will or have goofed the protocol. SciAm puts it like this:

Imagine standing at the laboratory bench, working on an experiment, when, as you finish one step, a display on the inside of your lab goggles tells you what to do next. A small camera in the frame watches your hands closely. If you reach for the wrong tube, the display flashes a warning. Before you can make the mistake, the system tells you how to get back on track.

Is that a total hoot or what?? Unbelievable. In my opinion, this is a big deal, not a little deal.

Q: Is Germaine biased and hyperventilating by this because it used the same kind of research as his was many years ago? 

🤨?

Footnote: 

1. p-Hacking happens when researchers analyze their data by what they thought was a good way to analyze it. But then, to their unhappiness, they find they missed statistical significance, usually expressed as a p or probability value. Simply put, if the data doesn't hit statistical significance the experiment and all the time and money that went into it led to was a failure. So, to try to get away from the failure, scientists analyze the data in another way in hopes, conscious or not, of getting to the magical probability value of p ≤ 0.05, or > 95% likely the result is real and not a statistical fluke.

AI glasses and an AI-driven robotic arm 

helping the human

By Germaine: Crusader for truth, justice and the American way . . . . . . also all for getting the replication crisis under better control