June, 2010SAS2010 Where does your child want to go today? Evgeny Kolotinsky / Content Filtering...

June, 2010 SAS2010

Where does your child want to go today?

Evgeny Kolotinsky / Content Filtering Technologies Development Group

Konstantin Ignatev / Web Content Analysts Group

June, 2010 SAS2010

• Nature of Internet dangers• New generation on the Internet• Techniques and examples

Points of interest

June, 2010 SAS2010

• Malware• Fraud • Phishing• Criminal activity

BUT:

are traditional research topics for information security companies.

The everyday evils of the Internet

June, 2010 SAS2010

Internet content itself can be dangerous

Especially for thosewho are too young or too inexperienced to

distinguish between good and bad.

Or for people who just want to spend a few hours at work doingmore interesting than just working.

June, 2010 SAS2010

The Nature of Internet Dangers

June, 2010 SAS2010

When we say “content is dangerous for kids” –

in most cases by “content” we mean “pornography”…

What content is dangerous?

June, 2010 SAS2010

“Pornography or porn is the portrayal of explicit sexual subject matter for the purposes of sexual excitement and erotic satisfaction.”

- Wikipedia

“I can't define pornography, but I know it when I see it.” - Justice Stewart in

Jacobellis v.Ohio 378 US 184 (1964)

What is it?

June, 2010 SAS2010

25 000 years ago

7 000 years ago

2 500 years ago

2 000 years ago

1 500 years ago

500 years ago

100 years ago

nowadays

Where it comes from

June, 2010 SAS2010

www.wikipedia.org

Pornography across the world

June, 2010 SAS2010

Online payment systems

Streaming content

Live chats

Broadband

Traffic optimization

Spam

Malware

Pop-ups

Browser hijacking

Domain-name hijacking

3G mobile services Paris Hilton

Pornography drives technology

June, 2010 SAS2010

Lena Soderberg (Sjooblom) from Playboy, November 1972

"Culture, Communication, and an Information Age Madonna," "IEEE Personal Communication Society Newsletter" Vol. 45, No. 3, May/June 2001

Pornography drives technology

– industry standard for digital image processing

June, 2010 SAS2010

Drugs & alcohol

It’s not just pornography

June, 2010 SAS2010

Violence & Extremism

It’s not just pornography

June, 2010 SAS2010

Gambling

It’s not just pornography

June, 2010 SAS2010

Wasting time

It’s not just pornography

June, 2010 SAS2010

Other4.4%

Uncategorized69.9%

Porn11.0%

Social networks3.9%

Chats2.1%

Shops6.5%

Gambling1.0%

Games0.8% Mail

0.4%

Hate speech0.9%

Violence0.5%

Payment sys-tems0.3%

Job search0.3%Weapons

0.1%Drugs0.1%

Illegal soft2.2%

www.alexa.com

Internet distribution, Top5000

June, 2010 SAS2010

The Internet Generation Gap

June, 2010 SAS2010

1981

1983

1984 1985

1989

1995 2001

2005

2010

The Internet is the greatest generation gap since rock and roll.

— Bruce Schneier

The Internet generation gap

June, 2010 SAS2010

www.emarketer.comwww.pewinternet.com

The Internet generation gap

2007 – 76%

2010 – 84.8%

Teens online

82% of teens help parents

52% of parents help teens

June, 2010 SAS2010

Parent’s Internet assumptions

We know something about where our children go and

what they do in the Internet

We have established rules for our child’s Internet

activity

88% 92%

www.isafe.org

June, 2010 SAS2010

Young peoples’ attitudes to the Internet

I do not share what I do or where I go on the

Internet with my parents

33%

I do not discuss Internet safety with my parents

40%My parents complain about the amount of time I spend on the

Internet

23%

My parents have no idea how much time I spend on the Internet

www.isafe.org

14%

My parents have not established any rules for

my Internet activity

34%

June, 2010 SAS2010

* Wolak et al, 2007

10-11 12-13 14-15 16-170

5

10

15

20

25

30

35

40

Unwanted and wanted exposure to online pornography among boys

WantedUnwanted

% o

f you

th in

tern

et u

sers

10-11 12-13 14-15 16-170

5

10

15

20

25

30

35

40

Unwanted and wanted exposure to online pornography among girls

WantedUnwanted

% o

f you

th in

tern

et u

sers

age age

The Internet generation gap

June, 2010 SAS2010

Our goal – protect against unwanted content

If someone wants to see adult content – nothing will stop him

• But we can make it more difficult• We should prevent adult content from being exposed in

other cases

• Parents should have the ability to protect their children from unwanted content

June, 2010 SAS2010

Content Filtering Techniques

June, 2010 SAS2010

Categorization Listing

Access and distribution

controlData filtering

Fundamental techniques

June, 2010 SAS2010

Local• Listings• Deep real-time analysis

Gateway• Listings• Real-time analysis

In The Cloud• Listings• Crawling• Expertise

Content filtering solutions

June, 2010 SAS2010

Pre-requisitions for content-filtering solution

• Listing means categorization• White list does not mean good, it just means not black

• All categorization criteria should be available• Categorization can be over-inclusive or under-inclusive

• Categorization does not mean blocking• Only the user can decide which categories should be

blocked

June, 2010 SAS2010

Kaspersky Parental Control

• Local solution (available in KIS 2009, 2010, 2011 etc)• Based on

• Listings• Heuristic analysis and categorization

• All categorization criteria available• Categorization is over-inclusive

• The user decides which categories should be blocked

June, 2010 SAS2010

Kaspersky Parental Control

June, 2010 SAS2010

Where does your child want to go?

Porn 72,2%

Illegal software

1.2%

Drugs1.8%

Violence3.8%

Hate speech6.6%

Weapons 3.7%

Gambling 7.0%

Chats0.2%

Web mail0.4%

Payment systems0.2%

Real FPs3.3%

• We can have false positives, and users can report them

• What are they ACTUALLY reporting as false positives?

* Kaspersky Lab, 2010

June, 2010 SAS2010

“I can't define pornography, but I know it when I see it.” - Justice Stewart in

Jacobellis v.Ohio 378 US 184 (1964)

Controversial categorization issues

June, 2010 SAS2010

Ooops…

Controversial categorization issues: example

June, 2010 SAS2010

Controversial categorization issues: example

June, 2010 SAS2010

Controversial categorization issues: example

June, 2010 SAS2010

Binary content analysis

Possible image analysis techniques

(Color, Shape, Similarity, OCR, etc)

At the moment, the current state of technology makes it almost impossible to categorize images

June, 2010 SAS2010

• Video analysis – frame by frame.

• Requires a huge amount of resources

At the moment, the current state of technology makes it almost impossible to categorize videos

Binary content analysis

June, 2010 SAS2010

Binaries

June, 2010 SAS2010

The Final Chapter

June, 2010 SAS2010

The Internet is dangerous, but it’s also an endless source of information

Any security content filtering solution is just a tool to help parents; it can’t replace parents

Children can try to find answers to their questions on the Internet or they can ask parents

Who will give the best answer?

It’s up to you!

June, 2010 SAS2010

Thank you! Questions?

Evgeny Kolotinsky, Konstantin Ignatev

ParentalControl@kaspersky.com