I also updated the text and appended ‘for now’. But clearly anything that humans can disagree on (which is a lot) is only that much harder for a computer. This stuff is hairy and no one is really talking about it in the mainstream world, let alone reasonable critique from technologists. Just read any social network analysis papers, and you’ll see how optimistic, “rational”, and simplistic most papers/authors are.

@zalet & steve – I had alt tags from the beginning, and plain text descriptions on the main page. I wish I had the time to create an alternative version for the blind that worked in sound, but alas I do not. If you wish to have access to the source code to do so yourself, I would be happy to collaborate.

Aaron

It’s an interesting project. The results I mentioned are exactly what you are trying to show.

I paid no attention to opening page and missed your point. I now see the new text on enter your name page.

How about something on the results page to emphasize some of the points you are making? “Is this you?” would seem to carry the point.

“Personas demonstrates the computer’s uncanny insights and its inadvertent errors, such as the mischaracterizations caused by the inability to separate data from multiple owners of the same name. It is meant for the viewer to reflect on our current and future world, where digital histories are as important if not more important than oral histories, and computational methods of condensing our digital traces are opaque and socially ignorant.”

As digital art, I think it’s very cool (nice job), which is why I originally used it to kick off my post. As philosophy, I also think the point it makes is similar to the point I’m trying to make, which is that the data that exists about us online is in many instances “independent from the true complexities and subtleties of reality.” Data mining/reality mining probably will someday fulfill the promise of the Tireless Machine Era, but as we start to see seeds of this, we will guard our online identities even more closely and manicure them ever more diligently.

But the real kicker is that much of what is uncovered in the Tireless Machine Era will still be unknown to us (and to others as well), but will be collected, aggregated, analyzed, monitored, and tagged by powerful, always-on systems running 24/7 in the background of all Web applications. The temporal aspect of one’s online profile – past actions, posts, connections, and data points – will be visible from the outside, creating trend lines that may not be evident to the individual user. To your point, the data mining will be opaque, yes; but socially ignorant, maybe not.

I’ll grant you that my only info about Personas was what I read in Naumi’s post and then a quick run of it after clicking through to the link, so you are correct that I read none of the other contextual background about the history that you provided – thanks for that.

But as to my other comments about the visual elements, I stand by them. Having served as the CEO of a major regional non-profit treating the disabled and having co-founded and seed-funded a non-profit that focused on making technology accessible to people with disabilities, I’ve heard comments like yours before and they are weak.

Many people go blind later in life, due to disease, injury, illness. By using accessible techniques to enable descriptive interpretation of an online work of art, you give someone who once had sight the ability to inwardly visualize it and, thus, call upon a level of appreciation for that beauty or creativity that they would otherwise not have.

You can learn more about just how to do this by reaching out to the Executive Director of Knowbility, Sharron Rush, at srush@knowbility.org. I’m sure she would be happy to help you with some tips on how to make Personas more accessible. Good luck!

Nice blog post. It is certainly true that much of our lives are being lived online, and therefore our digital identity can converge to a pretty good approximation of ourselves through our traces of history online. There is still a long way to go, but tools that aggregate and make sense out of our online histories will enable major new technologies to leap forward as computers have some kind of accurate social intelligence.

Now, to Doug Cornelius, Naumi Haque, and ESPECIALLY Steve Guengerich, a little background that’s missing on Personas. Did you guys realize that this is an art piece that is part of a larger exhibit now at the MIT Museum questioning our modern world (or even read what I wrote on the homepage)? It’s not a “tool” or a “proof of concept”. This is meant to critique data mining, and as such, include inaccuracies, models with fundamental assumptions baked in from the creator of the algorithms, only include partial kinds of data, be opaque in its final representation and how it arrived at its conclusions, etc. The viewer is meant to see the computer try to make sense of the data, oscillating back and forth and eventually converging on a box in which to put you in, no matter how appropriate or inappropriate, or with any explanation of its rationale. This is the world of data mining, and the “world of tomorrow”, where programmers and mathematicians attempt to create their models of how people interact and what data means independent from the true complexities and subtleties of reality… sometimes with success and sometimes not.

And besides, Steve– do you really think its appropriate for a research project done by a student at university (as a side project done in 2 months, nevertheless) to create programs for all languages in the world? Especially something that involves training large-scale language models? Come on. And besides, there were (and are) alt tags and meta data on the index.html for screen reading, and ultimately blind people can’t see a visualization so its not a terrible concern. And don’t forget, as art and not a tool, aesthetics matter more than attention paid to the color blind. I’d like to know if you complain about the colors on a Picasso. If you want to talk smack, at least have a valid critique that has some depth. You sound like a generic Media Lab hater: missing the big picture while arguing problems with things that were never the point.

Aaron

- what about the blind (no meta data or apparent alt tags in the result for screen reading),

- color blind (aren’t >15% of anglo men red/green color blind?), and

- non english speaking/reading

A classic Media Lab “proof of concept” good at stimulating ideas, but leaves a but to be desired regarding inclusiveness.

Some really interesting thoughts here.

The MIT personas is very interesting, but mis-characterizes itself. It does not produce a representation of person, it merely produces a representation of the name. I ended up with a big chunk labeled “sports” because it pulled information on Doug Cornelius, the basketball coach at Yuba College. (not me)

The other flaw is more central to your discussion. It only pulls information from the public internet. So it is missing the rich information in closed online communities. Facebook being the biggest missing piece. It also happens to be the place where people expose more of their personal side.

Different parts of the internet allow you to expose different parts of your persona.

In a professional space like this for you, or my blog for me, is probably of little interest to those who are near and dear. But in alternative spaces and closed online communities you publish information and communicate in a different way with different people than you would in a very public space.

I did a similar 4 box analysis pointing out that that the line is not merely personal versus professional, but also public versus private. You can see that here:
http://www.compliancebuilding.com/2009/04/13/the-4-ps-of-the-social-internet/