Google Declares War On Badware

(AP)  Google is phasing in a new feature in which it will issue this warning to people who try to click on links to sites with spyware and other malicious code: "The site you are about to visit may harm your computer!"

Users can search again, learn more about malicious code at the site StopBadware.org or proceed to the suspect site anyhow - at their own risk, of course.

Google Inc. said its initiative is just starting and is by no means comprehensive.

"To begin we'll only be identifying a small number of sites, but we'll be expanding our coverage over time," the company said in a statement. "Finding new and better ways to protect our users is a perpetual project, and we'll continue to work hard in this area."

Google is one of the main sponsors of StopBadware.org, a project that researchers from Harvard and Oxford universities are hoping to turn into a clearinghouse for information on spyware and other malicious software.

So far, StopBadware has identified only one site as malicious, and efforts to reach that site from Google worked normally Wednesday.

But Google has identified other sites as problems and is offering warnings for those. The company said the sites have been identified using software algorithms and verified with outside experts.

Jonathan Zittrain, a Harvard and Oxford professor involved with StopBadware, said the project may one day release lists of suspect sites still under investigation, but would most likely do so for everyone and not just sponsors like Google.
By Anick Jesdanun

[SGH2]

i meant numbers to show the degree of "danger." websteratcbs

[SGH2]

a way google could help would be to rate the types of "danger" by using something similar to the movie ratings. Use letters to symbolize nature of threat, and numbers to show perceived danger.

[bizologist]

VAPORWARE.

See the SiteAdvisor at Mc Afee - they already have it and it works great.

[hugbee]

Its about time someone did this. I've worked on search engines for 15 yrs and have a search engine patent. Years ago I mentioned to my peers that we can detect malicious code with a search engine. It would work the same way a virous scan works by matching patterns. It would be another tool in our corporate defenses.

The idea was never implemented because I was told it would never work and we don't need it. These are the same people who told me people do not want or need to be connected to the internet 24x7. And, "Why would you give customers access to their own data over the internet?". The stupid continues.

I'd like to congradulate Google for their forward thinking.

[alphaa10-2009]

This sounds like a good idea, provided there is general agreement on the technical basis for tangible risk. Some sites are Java-happy, others minimize Java. Likewise for the Active-X tribe-- both scripting routines have been used by hackers as a tool of choice. Perfectly honest and safe sites also use them.

So, Java or Active-X, by itself, will not fly as an indicator of risk. IE and Netscape provide optional controls for the degree of exposure to hacker Java or Active-X code, but they do not attempt to assess risk for a given page or website.

The article was short on details, but from other sources, it is clear Harvard and Oxford researchers have done a great deal of work on the problem. If their project could be made to work with the same reliability current third-party products (McAfee, Norton, etc.) claim on detecting risk from a given website, this would be a huge public service.

Now, if only the same researchers would turn their considerable talents to detecting the sources of spam.