Cell Phones Predict the Future

An anonymous reader writes "Wired News reports that cell phones were used in a recent project at MIT to both document and predict the lives of 100 MIT faculty and staff members. During the Reality Mining Project at MIT, Researcher Nathan Eagle logged 350,000 hours of data over nine months about the location, proximity, activity and communication of volunteers through cell phones carried by the participants. From the article, "Given enough data, Eagle's algorithms were able to predict what people -- especially professors and Media Lab employees -- would do next and be right up to 85 percent of the time."

Comments

I predict that data thieves will love this!

[garcia]

"We want to have our life choreographed, cataloged, witnessed and archived," Stakutis said. "Now we are heading to a world where this is possible without effort."

Do we? It's one thing to have a personal diary or blog that you opt-in to submit information to daily. Hell, I have even expanded on my mobile pics to include a "blog" of what I did during any particular day... That's my *choice* to put that information out there for people to see. It's not mandated by my cell phone to take pictures of what I'm doing and throw them into a database that I have no control over.

While Eagle "acknowledges that the project raises some important questions about privacy and about the ownership of data, and says people should feel empowered, not scared, by his cell-phone applications," I just can't get passed his statement earlier in the article:

The Media Lab behavior is beautifully regular, but the lab lives and dies by sponsors' meetings," Eagle said. "So the weeks leading up to sponsors' meetings, people are pulling all-nighters and people are going crazy trying to get their demo working.

Is this another demo for one of your sponsors that might end up buying the rights of this technology from you and then creating their own spyware network of their mobile users' daily habits? Tracking when, where, and how they communicate to "better" serve them with advertisements and the selling/stealing of their data to other institutions and data thieves?

He has already founded a company called MetroSpark that in September will launch a Bluetooth-powered social-introduction service.
After filling out a personal profile, MetroSpark will attempt to be a gracious, ubiquitous host that connects people with common interests, whether they are technology conference goers who share an interest in motorcycles or barhopping singles who love long walks on the beach at sunset.

Oh, so you started this company -- got it advertised on Wired and now Slashdot -- and it's never going to get bought out by someone else (i.e. Dodgeball) and they aren't going to use this huge database of customer data that was originally meant to be benign?

I predict that even more corporations are going to have a field day with this data than what they originally intended (i.e. when/where you have your cell phone on and how many days a week you are sitting at home letting the CATV wash over you). If the corporations (and then obviously the government) can track social networks and trends via software on the phones you can bet your ass they are going to include it "free of charge" while still restricting your "free" access to any other programs you might want to run.

I predict that people will fall for this invasion just like any other. We're seriously one step closer to the "Big Brother" that everyone used to fear... Now we are welcoming him with open arms!

Re:I predict that data thieves will love this!

[HTH NE1]

"We want to have our life choreographed, cataloged, witnessed and archived," Stakutis said. "Now we are heading to a world where this is possible without effort."

Indeed, next comes the government contract to expand and fully exploit this information. Soon, local law enforcement will be using this data to do their jobs more efficiently and stopping people for questioning just because they've "strayed from the herd".

And they'll do it without directly violating your privacy because they won't see the data that was the basis of the alert. As long as no one but the black box doing the mining sees your private information and doesn't disclose any of it with its findings, it's not going to be seen as a violation of your privacy. Privacy violations will become defined as disclosure of one person's information to another person, and machines running automated processes will be exempt by definition.

Shameless Plug - Schedule Nanny

[metlin]

Hmm, not entirely the same thing, but I'd worked on a project called ScheduleNanny, where we used people's PDAs coupled with GPSes to predict where they will be.

There were some interesting emergent behaviors - for instance, the system would know that I have to go to the bank later in the day and I would drive by the bank in the morning, so it would indicate that I could save time by going to the bank then. Or for instance, it would beep in the morning that it was time for me to go shower or go to the train station.

Details can be found here.

All in all, it was pretty good - after some amount of initial bootload information, you can take away the GPS and quite accurately predict where people are likely to be. This looks fairly similar, in some ways.

Which means no predection at all

[Anonymous Coward]

In metheorology it is a fact, that if you predict the next day weather to be excactly the same that it is today, you end up with 85% average.

Well in college I'm usual in one place.

[teiresias]

Given enough data, Eagle's algorithms were able to predict what people -- especially professors and Media Lab employees -- would do next and be right up to 85 percent of the time."

Course, in my college days, if my cell phone predicted I'd be in the computer lab, 99% of the time it'd be right.

My cell phone is telling me...

[ucahg]

My cell phone is telling me that on thursday I will read this story again.

Sounds UNimpressive to me...

[dpbsmith]

I can predict that the next time I weigh myself the scale will read between 160 and 170. This prediction would have been true far more than 85% of the time over the last five years and I will be very surprised if it is not true the next time I weigh myself.

Once I learn that someone works a full-time job and where they work, I can predict with greater than 85% accuracy where they will be between 10 a.m. and 2 p.m. on Monday through Friday.

I've heard it said, whether or not correctly I do not know, that if you simply predict that tomorrow's weather will be the same as today's, you will be accurate more often than the weather service.

Predictions are only valuable when they are unlikely or surprising. Tabulating obvious patterns and predicting their continuation may be highly accurate yet low in value.

Old News

[asscroft]

My cell phone told me this yesterday!