SETI@home a security risk?

December 8, 2005 | Astronomy | By: Mark VandeWettering

This rather wacky report talks about the possibility that ETs could infect the earth’s computers with viruses by transmitting certain signals which cause (say) buffer overflows in our computers and infect the internet.

Dr. Carrigan thinks that the SETI scientists should implement some kind of decontamination procedure to clean the signals before they are distributed.

It seems kind of obvious to me that if we knew enough to conclude that a particular signal was trying to act as a virus, we’d have pretty strong proof that extra-terrestrial intelligence was a reality.

I tried to locate the actual article, but could only find this link to the abstract.



Comment from Theo Honohan
Time 12/8/2005 at 10:57 am

He’s an idiot. You can get some articles from his website, here:

It’s pretty clear that he is unable to reason clearly about the distinction between code and data, and doesn’t understand in detail the way viruses, worms and buffer overflows work in practice. The biological analogy is broken because there is no code/data distinction in nature… it’s all molecules and atoms.

Comment from Dan Lyke
Time 12/8/2005 at 12:51 pm

Because, presumably, he’s concerned that the aliens might find an exploitable bug in the either SETI or associated network code and broadcast a sequence that…

I’ve got a better idea: To do this the aliens would have to get the SETI code. I don’t know how often upgrades to the code are released, but we’ll say that you’ve got two years to find a hole and exploit it. And it seems like we know what’s going on in space to say a light minute or so around the earth. So if we scan net for ping times between two minutes and a year or two, we should be able to find the alien’s IP addresses and simply disallow downloads of the SETI software to them.

Comment from turchin
Time 10/4/2007 at 4:23 am

Is SETI dangerous?

If we think that Seed AI is possible, then SETI is much more dangerous, then it was thought by Richard Carrigan in his article ‘Do potential Seti signals need to be decontaminated?’

Any supercivilisation, that reach AI, could send in space pieces of its own source code by radio. It will be like space viruses or trojans that will be executed by naïve early civilizations and replace them. Of course it will be not only plain code. A simple and clear method is exist how to send through radio signals a program which will be compatible with any computer. Here is needed steps:

1. At first it will contain video, where extraterrestrial will explain their language and promise a lot of benefits, if the code will be executed. (Video will be easily recognized by signals of line end and picture end, as it was in old TV systems on Earth)

2. Then the video will explain how to built simplest possible Turing computer. (i.e. Post computer, which has only 6 commands, and is so simple that thereis no alternatives to it. One of the way of that explanation is to show details of it – diodes, transistors, and whole scheme.)

3. Then they will send a program for Turing computer, that will contain a code of Seed AI, capable for learning but steady in his goals.

The main problem here is that will be that many states and organizations on Earth which will know specific point from space from which intellectual signal is coming. So there will be competition – who first run the code and get all promised advantages. We have risk that at least one group decided that they could control the extraterrestrial code and run it.

As we have more and more powerful radiotelescopes, the possibility that we encounter extraterrestrial radio signal is growing and I think a large part of all extraterrestrial signals could contain vicious code, because self replicating is only good reason to send a lot of information in space and spend on it a lot of energy. And it is a quickest way for supercivilisation to expand in the speed of light.

Alexey Turchin,