The calculator paradox

When was the last time you checked the result of a calculator? Can anyone still imagine that people used to calculate “by hand”?

For decades, we have been brought up to trust technology. First the calculator, then computers, now smartphones and tablets. The art of engineering has produced some truly insane and crazy devices. Technology is calculable, deterministic, comprehensible and – conversely – predictable.

Our generation has become accustomed to trusting technology, we have also stopped checking the results of calculators with pen and paper and we often trust computers more than the users who operate them.

๐—จ๐—ป๐—ฑ ๐—ท๐—ฒ๐˜๐˜‡๐˜ ๐—ธ๐—ผ๐—บ๐—บ๐˜ ๐—ž๐—œ – ๐˜‚๐—ป๐—ฑ ๐—ฎ๐˜‚๐—ณ ๐—ฒ๐—ถ๐—ป๐—บ๐—ฎ๐—น ๐—ธ๐—น๐—ฎ๐—ฝ๐—ฝ๐˜ ๐—ฑ๐—ฎ๐˜€ ๐˜€๐—ผ ๐—ป๐—ถ๐—ฐ๐—ต๐˜ ๐—บ๐—ฒ๐—ต๐—ฟ… !

Due to the structure of AI (neural networks, training data, weights) and the incredibly large amounts of data, systems are created that are no longer comprehensible to humans (or other machines). AI cannot be explained or calculated. The consequences are unforeseeable and will keep us busy for years to come.

In future, our society will no longer be able to rely on the familiar trust in technology. AI systems must be categorized differently and their results evaluated differently. This is a huge challenge in teaching at the PFH Private University of Applied Sciences Gรถttingen and also in the field of experts ( Streitz Hoppen & Partner – IT experts ). But: we are facing up to this challenge, because we must not “rely” on the fact that it will somehow fit.

We are currently in a phase of fascination and, above all, in a phase in which the users of AI can still judge for themselves whether the results make sense or not. We (still) have the expertise to be able to assess and evaluate the results.

๐—”๐—ฏ๐—ฒ๐—ฟ ๐˜„๐—ฎ๐˜€ ๐—ถ๐˜€๐˜ ๐—ถ๐—ป ๐Ÿฑ ๐—๐—ฎ๐—ต๐—ฟ๐—ฒ๐—ป? ๐—ช๐—ฎ๐˜€ ๐—ฝ๐—ฎ๐˜€๐˜€๐—ถ๐—ฒ๐—ฟ๐˜, ๐˜„๐—ฒ๐—ป๐—ป ๐˜„๐—ถ๐—ฟ ๐˜€๐—ฒ๐—น๐—ฏ๐˜€๐˜ ๐—ป๐—ถ๐—ฐ๐—ต๐˜ ๐—บ๐—ฒ๐—ต๐—ฟ ๐—ฏ๐—ฒ๐˜‚๐—ฟ๐˜๐—ฒ๐—ถ๐—น๐—ฒ๐—ป ๐—ธ๐—ผฬˆ๐—ป๐—ป๐—ฒ๐—ป, ๐—ผ๐—ฏ ๐—ฒ๐—ถ๐—ป ๐—˜๐—ฟ๐—ด๐—ฒ๐—ฏ๐—ป๐—ถ๐˜€ ๐—ฆ๐—ถ๐—ป๐—ป ๐—บ๐—ฎ๐—ฐ๐—ต๐˜, ๐˜„๐—ฒ๐—ถ๐—น ๐˜„๐—ถ๐—ฟ – ๐—ถ๐—บ ๐˜‚ฬˆ๐—ฏ๐—ฒ๐—ฟ๐˜๐—ฟ๐—ฎ๐—ด๐—ฒ๐—ป๐—ฒ๐—ป ๐—ฆ๐—ถ๐—ป๐—ป – ๐˜ƒ๐—ฒ๐—ฟ๐—น๐—ฒ๐—ฟ๐—ป๐˜ ๐—ต๐—ฎ๐—ฏ๐—ฒ๐—ป, ๐—บ๐—ถ๐˜ ๐—ญ๐—ฒ๐˜๐˜๐—ฒ๐—น ๐˜‚๐—ป๐—ฑ ๐—ฆ๐˜๐—ถ๐—ณ๐˜ ๐˜‡๐˜‚ ๐—ฟ๐—ฒ๐—ฐ๐—ต๐—ป๐—ฒ๐—ป?

Perhaps we should also check our calculators… because none of us would notice if the calculation shown in the picture produced an incorrect result.

๐—•๐—น๐—ถ๐—ป๐—ฑ๐—ฒ๐˜€ ๐—ฉ๐—ฒ๐—ฟ๐˜๐—ฟ๐—ฎ๐˜‚๐—ฒ๐—ป ๐—ถ๐—ป ๐—ง๐—ฒ๐—ฐ๐—ต๐—ป๐—ถ๐—ธ ๐—ถ๐˜€๐˜ ๐—ด๐—ฒ๐—ณ๐—ฎฬˆ๐—ต๐—ฟ๐—น๐—ถ๐—ฐ๐—ต – ๐˜‚๐—ป๐—ฑ ๐—ถ๐—บ ๐—ž๐—œ-๐—ž๐—ผ๐—ป๐˜๐—ฒ๐˜…๐˜ ๐—ฒ๐—ฟ๐˜€๐˜ ๐—ฟ๐—ฒ๐—ฐ๐—ต๐˜…

The paradox: we trust a calculator more than we trust ourselves… crazy.

#informatikersindcool#kiistdaundwirdbleiben