johnson slides 2020 final · horizon scanning using machine learning refugee flows and instability*...
TRANSCRIPT
Future Crime
Professor Shane D JohnsonDawes Centre for Future Crime at UCL
Dawes Centre for Future Crime at UCL
Office for National Statistics
Dawes Centre for Future Crime at UCL
0.00
5000.00
10000.00
15000.00
20000.00
25000.00
Dec
81
Dec
87
Dec
93
Dec
97
Mar
02
Mar
04
Mar
06
Mar
08
Mar
10
Mar
12
Mar
14
Mar
16
Jun
18
Thou
sand
s of o
ffenc
esAll CSEW crime
Exc Fraud andComputer Misuse
Inc Fraud andComputer Misuse
Office for National Statistics
Dawes Centre for Future Crime at UCL
0.00
5000.00
10000.00
15000.00
20000.00
25000.00
Dec
81
Dec
87
Dec
93
Dec
97
Mar
02
Mar
04
Mar
06
Mar
08
Mar
10
Mar
12
Mar
14
Mar
16
Jun
18
Thou
sand
s of o
ffenc
esAll CSEW crime
Exc Fraud andComputer Misuse
Inc Fraud andComputer Misuse
Dawes Centre for Future Crime at UCL
A classic Crime Harvest and Retrofit SolutionDawes Centre for Future Crime at UCL
There were only two permutations!
+ = Crime opportunity
+ = No crime opportunity
Future Crime
• How do we avoid crime harvests?
• There are no “future facts” Jouvenel (1967)
• Challenge of methods for futures….
– Horizon scanning– Delphi– Stated preferences– Penn testing– SRs
• Where to start?
Dawes Centre for Future Crime at UCL
Addressing Online CSE on Social Media
Crime, place and the internet
The effects of cyber weapons
Smart Cities: Opportunities for crime prevention
Cybercrime risks to future street infrastructure
Guarding against Adversarial perturbation
Biocrime
Detecting emerging crime in
online markets
Horizon Scanning using Machine
Learning
Refugee flows and instability*
Current Research Themes
Terrorist financing and money laundering
Crime, place and the internet Cryptocurrency &
Fraud
Advanced materials to
combat crime
Recent and future trends in
counterfeit goods
Developing technologies
IoT and Crime
Refugee flows and instability*
Scoping projects
Longer projects
PhD projects
Dawes Centre for Future Crime at UCL
Ageing and online fraud
AI and Crime
Reducing domestic abuse with technology
Consumer Internet of Things (IoT)
Source: Intel
Dawes Centre for Future Crime at UCL
Consumer IoT
Race to market + No security regulation + Cost = ?
Dawes Centre for Future Crime at UCL
Systematic Review of Crime Facilitated by IoT
Initial Search3506 Initial
Screen2708Full
Screen198 Coded114
Databases Searched: Web of Science, ProQuest, ACM Digital Library, IEEE Xplore Digital Library, and Scopus
Dawes Centre for Future Crime at UCL
Blythe, J. M., & Johnson, S. D. (2019). A systematic review of crime facilitated by the consumer Internet of Things. Security Journal, 1-29.
Thematic Synthesis
Dawes Centre for Future Crime at UCL
Methods of Attack
Method Description
Denial of Service (DoS) Services disrupted through > requests from single devicesDDoS DoS attacks from multiple sources (e.g. botnets)Malware Malicious s/ware to compromise device and change functionsMan-in-the-Middle Interception of data in transitPhysical attacks Attacker physically tampers with deviceData integrity Attacker compromises data — inserting, modifying, replay attackSpoofing Attacker masquerades as anotherSide channel Exploits seemingly non-sensitive info e.g. power usageUser impersonation Attacker impersonates user (e.g. social engineering attack)
Dawes Centre for Future Crime at UCL
Methodology & Plausability
Dawes Centre for Future Crime at UCL
Crimes/Harms
Dawes Centre for Future Crime at UCL
Dawes Centre for Future Crime at UCL
Open Source Search
• News media published in two 60-day periods: – 20 January 2018 to 21 March 2018, and – 14 May to 13 August 2019
• NLP screening
• Google searches (same period, 99 articles)
Initial Search64k NLP
Screen1.6k FullScreen287
Dawes Centre for Future Crime at UCL
Crimes/Harms
Dawes Centre for Future Crime at UCL
Secure by Design
• Warning signs of a crime harvest
• Heterogeneity – can you tell if a device is secure before you buy it?
Dawes Centre for Future Crime at UCL
Smart Cities
A city that uses information and communication technologies (ICTs) and all other technologies available to improve the effectiveness and efficiency of city services in order to save resources and improve citizens’ well-being
Ramaprasad et al. (2017)
Laufs, J., Borrion, H., and Bradford, B. (2020). Security and the smart city: A systematic review. Sustainable Cities and Society, 55
Dawes Centre for Future Crime at UCL
Smart Cities
Laufs, J., Borrion, H., and Bradford, B. (2020). Security and the smart city: A systematic review. Sustainable Cities and Society, 55
Dawes Centre for Future Crime at UCL
Laufs, J., Borrion, H., and Bradford, B. (2020). Security and the smart city: A systematic review. Sustainable Cities and Society, 55
Dawes Centre for Future Crime at UCL
Smart Cities
Laufs, J., Borrion, H., and Bradford, B. (2020). Security and the smart city: A systematic review. Sustainable Cities and Society, 55
Dawes Centre for Future Crime at UCL
Smart Cities
Laufs, J., Borrion, H., and Bradford, B. (2020). Security and the smart city: A systematic review. Sustainable Cities and Society, 55
Dawes Centre for Future Crime at UCL
Smart Cities
Laufs, J., Borrion, H., and Bradford, B. (2020). Security and the smart city: A systematic review. Sustainable Cities and Society, 55
Dawes Centre for Future Crime at UCL
Systematic Review – Systematic Approach?
• Crime/Security are often omitted from (Smart) city plans
• Needs a systematic approach - dependencies across (silo) systems and sectors
• Governance:– Who owns the data?– Balancing collective security and individual privacy?
Dawes Centre for Future Crime at UCL
Artificial Intelligence
Dawes Centre for Future Crime at UCL
Machines are learning to find concealed weapons in X-ray scansCOMPASS@CS_UCL
https://www.economist.com/news/science-and-technology/21711016-artificial-intelligence-moves-security-scanning-machines-are-learning-find
http://www.economist.com/news/science-and-technology/21711000-week-how-optical-navigation-can-help-bomb-find-its-target-without-gps
Autonomous software
AI to prevent crime
Dawes Centre for Future Crime at UCL
Adversarial Perturbations (speech to text)
Carlini, N., and Wagner, D. (2018). Audio Adversarial Examples: Targeted Attacks on Speech to Text. Deep Learning and Security Workshop.
Dawes Centre for Future Crime at UCL
Google AI thinks rifle is a helicopter (Can you spot the difference?)Adversarial examples are easier to create than previously understood
MIT team reliably fooled Google’s Cloud Vision API, a machine learning algorithm used in the real world today.
http://www.labsix.org/physical-objects-that-fool-neural-nets/https://www.wired.com/story/researcher-fooled-a-google-ai-into-thinking-a-rifle-was-a-helicopter/
Defeat to AI
AI to prevent crime
AI to commit crime
Adversarial perturbations
Dawes Centre for Future Crime at UCL
Generative Adversarial Networks (GANs)
https://forms.gle/9vD5hHHKNtc7y5p3A
Dawes Centre for Future Crime at UCL
Dawes Centre for Future Crime at UCL
AI and Crime
Dawes Centre for Future Crime at UCL
Dawes Centre for Future Crime at UCL
Biotechnology
• Traditional biological systems are re-created or modified in novel ways for various application purposes
• CRISPR, Gene drives, DNA testing
• In 2018, £2.2bn was raised by biotech companies in the UK
Dawes Centre for Future Crime at UCL
Synthetic Biology
• Biohacking - Athletes have historically abused steroids and growth hormones (Reardon & Creado, 2014)
• 23AndMe data sharing - informed consent?
• DNA encoded malware (Ney et al., 2017)
Dawes Centre for Future Crime at UCL
“Now that DNA sequencing, synthesis, manipulation, and storage are increasingly digitized, there are more ways than ever for nefarious agents both inside and outside of the community to compromise security.”
Peccoud et al 2017
Dawes Centre for Future Crime at UCL
Addressing Online CSE on Social Media
Crime, place and the internet
The effects of cyber weapons
Smart Cities: Opportunities for crime prevention
Cybercrime risks to future street infrastructure
Guarding against Adversarial perturbation
Biocrime
Detecting emerging crime in
online markets
Horizon Scanning using Machine
Learning
Refugee flows and instability*
Current Research Themes
Terrorist financing and money laundering
Crime, place and the internet Cryptocurrency &
Fraud
Advanced materials to
combat crime
Recent and future trends in
counterfeit goods
Developing technologies
IoT and Crime
Refugee flows and instability*
Scoping projects
Longer projects
PhD projects
Dawes Centre for Future Crime at UCL
Ageing and online fraud
AI and Crime
Reducing domestic abuse with technology
And finally
• Innovation-crime sequences
• Multidisciplinarity (AI, synbio, chemistry, sensors, cyber, ……)
• Multi-stakeholder (Gov, police, voluntary sector, industry, ……)