Earlier this week the White House unveiled an online system to submit questions to the President and vote on other people submitted questions. My first reaction was very positive. I liked the fact that I discovered this as a post in the stream on my FaceBook page. I was eager to see what kind of technology they would be using for the system. I have been impressed with the WhiteHouse.Gov web site team’s work ever since they switched it on at exactly 12:00pm on January 20th.
I navigated to the site and was pleased to see it was very responsive and handling what had to be an incredible number of users. I would like to see their roll out plan and how they ensured it’s scalability. I signed up for an account and submitted my question in about 5 min. My question was, verbatim:
How will you prevent gaming of this system? I imagine there are special interests with the technology to get their agenda onto this system – submitting hundreds of votes for the questions they want to appear important etc
In a matter of hours my question received 10 negative votes and 7 positive votes. It didn’t come close to the 5k+ votes that the most popular questions received.
As it turns out, the people at NORML did exactly what I thought would be done. They rallied their large following online to game the system, and all of the most popular questions in several categories were questions about the legalization of marijuana.
So I’m a bit disappointed by the White House technology team – they should have seen something like this coming. I don’t know if there’s an iron clad method to prevent this. I don’t think it’s even so much a technology problem. I am also not faulting NORML in this – as they did exactly what they were supposed to do.
The only idea I have that might help is less anonymity in the system. In this particular instance it might have prevented people from publicly stating their position on the marijuana issue – but more importantly, if you can tie the voting system to real people you can prevent the “multiple votes” problem, which I suspect was the case here.
There is some technology that might help as well, like the fully open online voting system called Helios. It’s a system that uses pretty standard computer security techniques to enable people to submit votes secretly but still verify that their vote was used in the final tally. I don’t propose that such a system be used to the point where every person who votes does the math to verify their vote was counted – but the existence of such a technique in the system would ensure that someone could audit the election, and anyone questioning the validity could do so.
Helios doesn’t solve the “one vote per person” problem, but I believe it could be used in conjunction with some other system to do so. What that other system will look like I have some ideas about – but it opens up privacy and indenty theft issues that mass media has convinced people are more important than things like participation in government.