Handwriting Software: Risk Vs Reward

[Logethica]

New Computer Software Can Copy Your Handwriting:

Researchers from UCL, University College London have developed a computer software program called, ‘My Text in Your Handwriting,’ which uses a person’s handwriting sample, accurately replicating it....



Published in ACM Transactions on Graphics and funded by the EPSRC, the machine-learning algorithm is built around glyphs – which analyses handwriting, reproducing it in the style of the sample.

How ‘My Text in Your Handwriting’ Works

The researchers asked a user to write on four A4-sized sheets of paper. This was then scanned and interpreted by a computer and human moderator, examining the position of letters and the style words were written in.

The team then asked a group to decide if any envelopes (which were written by the computer software) were created by a computer. In 40 percent of cases, they were incorrect.

The software could help in detecting forged documents, improving identity theft and online security.

With all the positive benefits, some could also use this computer software to forge signatures and handwriting for criminal and personal gain..

[To read the full article please visit the link at the top of the page]​

MT Members: Do you feel that this software is an advantage or disadvantage in minimising Signature & Handwriting fraud?

 

[Cohen]

I feel like this could be very dangerous if it was in the wrong hands. I'd probably feel safer if this technology wasn't used very much at all.

 

[DardiM]

Thanks for the share

I voted " I am unsure".
Because every tool can finally be used to do bad or good things, depending on who use this tool.

I feel like this could be very dangerous if it was in the wrong hands. I'd probably feel safer if this technology wasn't used very much at all.

I totally agree with this part

"The team then asked a group to decide if any envelopes (which were written by the computer software) were created by a computer. In 40 percent of cases, they were incorrect".
It would be interesting to know exactly the number of person. And how they "selected" them.
Because I think 40 % from some groups could not represent a validation