Harvard Prof Says Computers Need to Forget

Jessamine writes "A Harvard professor argues that too much information is being retained by computers, and the machines need to learn how to forget things as humans always have. "If whatever we do can be held against us years later, if all our impulsive comments are preserved, they can easily be combined into a composite picture of ourselves," he writes in the paper. "Afraid how our words and actions may be perceived years later and taken out of context, the lack of forgetting may prompt us to speak less freely and openly." Will such massive databases make us all act like politicians? Is data retention creating a "panopticon"? These are questions that the good doctor raises."

Comments

No kidding

[wiredog]

I just pulled my credit report, which I do 3 or 4 times a year. My address at my old air defense outfit at Ft Polk (D 1/55 ADA) is still listed there.

I got out of the army in 1988.

Public databases need to forget

[GauteL]

Right now, you can find 10 year old newsgroup postings from myself when searching Google. That is in my opinion too much data retention. I don't think they are particularly embarrassing now, but there was a time when my 10 year old newsgroup postings would have been posted when I was 14-15 or so.

They embarrasseed me, and I don't think I should be expected to be cautious about this issue when in my teens. People forget most of your silly mistakes from such a while ago but databases do not, unless you instruct them to.

Unrealistic

[Anonymous Brave Guy]

I'm all for personal responsibility, but this isn't as simple as standing by what you once said.

For one thing, no-one is perfect. If I took offence every time one of my friends said or did something a little childish or hurtful to another of my friends, I would have few friends left, yet I know that all of my friends are basically nice, decent people, who on balance I am glad I met. Magnify this up to the whole world stage, and suddenly the whole world is an a**hole.

Secondly, people's views change for many reasons, not all of them bad, and society as a whole is not good at recognising this. Just look at what happens to politicians today who change their position on an issue. "U-turn! U-turn!" As I've pointed out before, even in politics it is silly to think that our elected representatives have the time to fully study each issue on which they vote in the same detail as an expert, or to retain a staff of suitably smart and qualified people who can at least advise them well. Wouldn't you rather be represented by someone who would change their mind if they realised their previous position was short-sighted or ill-informed, rather than one forced by the system to stick to their guns even if they knew they had made a mistake?

I've commented on this subject before on Slashdot, in the context of social networking sites. I think humanity needs to learn that in a highly-connected world, you have to be careful what you say, you have to be wary of reading too much into what others say, and most important of all, you have to cut people a little slack sometimes. Right now, IMHO, our laws don't place nearly enough value on privacy, and I think this is a painful lesson that we are going to learn as an entire generation who grew up with the likes of Facebook, Myspace and LiveJournal run into problems for the next few decades.

Bottom line: kids will be kids, adults will behave like kids sometimes, even the most mature and responsible adult makes mistakes, and all of this is only human. I, for one, would prefer not to live in a world where everyone's dirty laundry was aired in public, with full search features.

Polluting the well

[HangingChad]

Yes, it's true specific incidents can last seemingly forever on the internet. But when you look at the increase in available data, it's staggering. And that data is spread out around hundreds of servers in different formats. But the more data, the more junk data. Outdated, incorrect entries and, sometimes, forged entries.

I manage a lot of data and have learned over the years how easy it is to pollute that well. Users are ingenious in their ability to get crap information in a system, no matter how tight you think your validation is. And importing data from an outside source...lol...even more of a nightmare. The ability of the internet to store information long term can also be used to hide information by clouding the waters.

So, years ago I started polluting my personal online data well. Instead of one or two profiles, I'd have five or six, all different. Different addresses, phone numbers, cities, states even race and gender. Five turned into ten, turned into 15 or 20 and then I lost count. Started doing the same thing to my online resumes. Cloned the resume under a different name, address and phone number. Created new resumes with the name withheld or changed with many subtle variations, swapped out phone numbers, email addresses. Started masking personal information behind my own LLC. Turned into my own personal Wild Weasel. Which one of the clones is really me? Hard to tell.

Haven't done that to my direct marketing profile and credit report...yet. But the day may come when I want to poison those wells. Don't need instant credit, pay cash for almost everything, including cars. You can play hell with your credit report by getting a camper or a boat and living on one of them for a while. I could park in my nephews back 40 for a couple months. Or live overseas, almost as good. Use PO boxes, change addresses so many times no one can keep up. Have one address for drivers license and vehicle registrations, a different one for online orders, another one for tax purposes. Use the wife's cell phone one week, mine the next. Change phone numbers twice a year. And, ironically, it's computers that give me the ability to keep up with all the different versions of myself.

No matter how good you are at consolidating data, there's always going to be someone like me with the knowledge to crap it up and make you work at manual consolidation. Got a lot of spare time on your hands to figure that out? :)