HEAR BEATE LIVE at Ideas on October 20, Los Angeles 2018

Click here to get your ticket for the must-attend event Oct. 20 for anyone that is already in Data Science or just looking to expand their career network.

In response to my blog post My New Fake Artist Portfolio I am posting the response from the programmer. While at first I was ready to be very upset with Sures Kumar – I must say that after reading his email to me I believe he is making a very serious point, something I have not thought about before. The issue we are facing is less about this one project he programmed on creating fake artist porfolios, but about how easy it is to create fake profiles, fake people, fake portfolios, fake anything. Take a read and let me know what you think…  What’s the big issue in your opinion?

Hi Beate Chelette,

Read your post reg Pro-Folio and here is my response to it.

The idea is not to steal and make money or create real people’s portfolio’s. The portfolios created through this website are completely fictional even though it has the people’s name on it. You must notice in “about” page, the email id generated and the description of that person is completely random and the person trying to create this fake portfolio has no control over it. Also every time you enter a person’s url like http://www.pro-folio.org/steve-jobs/ it creates a new set of random works. Note: all the portfolio details generated temporarily are stored in the form of cookies in your local computer. This was intentional. So when you clear your browser’s cookie, you will see a diff set of works. Only if you copy paste the exact url generated after clicking on a project, you can replicate the same page again. which is what Mark Wilson has done in his article. Even this newly created Mark Wilson is a fictional character with the name ‘Mark Wilson’ who was born in Nigeria, studies in Stanford blah blah. Real Mark is not taking credit of any of the works.

Also I would like to make it clear that I am in no way making it easier for someone to use artist’s images. The created fictional identities are of no use to anyone because expect the name all other details like email, country, education profile wont match with the person who is trying to use this fictional portfolio.

Currently the works “images” taken from Behance is only used to construct fictional characters profiles with your name on it. So theoretically that person does not exist. Its a cyber machine generated identity. This was intentionally done to make sure people don’t misuse Pro-Folio.

The idea is not to prove that someone can use an artist’s image without permission. The idea is to show that a machine (few lines of code) can pick up information from random sources, make sense of it and mash it up in a meaningful way to fool us “humans”. This is more than just copy-pasting the artist’s images. Imagine tomorrow if there are 300 Beate Chelette’s portfolio online created by various such bots, how do you think your audience can differentiate the machine generated ones from your original one in search results. I want my audience to think about this.

Through this project, I would like the audience to think about machine-generated identities and their influence in our society. Today they exist in many forms like Physical Robots, Intelligent speech engines (Siri, Google voice), and Automated chats in customer support. We also experience the crude version of such identities when we get that spam email asking to transfer money to an off shore account. With the rapid development in computing, there will be more such machine generated human identities in the coming days. There will be a point where we wont be able to differentiate a real human identity from a machine generated one.

By creating a model prototype of such identity creating machine, pro-folio also aims to question possible motivations to create such identities. Also hints at the possible future where machines in no time can do such a thing. If so, who will be interested in doing this? Hollywood? Private Organisations? Gaming industry?  and questions “what sort of systems do we have in place to differentiate a real human from a machine”. Can we stop large scale organisations from doing this? Theoretically anyone with the right infrastructure can scrap the entire internet for intelligence to come up with extremely believable identities. My intention is to raise such questions among the audiences and possibly encourage discussion.

Hope you found this information useful.

Have a good day!

Regards,
Sures Kumar

Pin It on Pinterest

Share This