AI summary by TwinMind
Can AI Catch Criminals at Sea? | Dyhia Belhabib | TED
TED
10m 54s
Speaker Background & Ocean Crime Investigation
- Scientist investigating ocean crime and behavior
- Cannot swim, born in landlocked Algerian village
- Investigated Spanish-owned vessel "Asian Warrior" in Dakar, Senegal
- Vessel on Interpol radar for illegal fishing
- Escaped Thai custody under name "Tai Chan"
- Observed modern slavery indicators on vessel:
- Crew eating only rice on floor
- Cardboard mattresses, no blankets
- Low ceiling sleeping quarters
- Rust and dead fish in gray water everywhere
Scope of Ocean Crimes
- Counted nearly 50 different types of ocean crimes:
- Drunk driving/navigating
- Illegal fishing
- Drug trafficking
- Wildlife smuggling
- Terrorism
- Murder
- Most crimes happen out of sight of authorities
- Example: Video of people being shot while swimming in ocean
- Only discovered because perpetrator forgot phone in Fiji cab
- Cab driver gave it to authorities
Ocean Crime Statistics & Visibility
- Oceans cover 2/3 of planet surface
- Nearly 4.6 million vessels worldwide
- Only 2% visibility into vessel activity
- Lack of visibility exploited by criminals for illicit operations
Impact on Land-Based Communities
Vancouver Real Estate
- Maritime drug trafficking money laundering increased real estate costs
- Chinese Triads injected over $4 billion into Vancouver real estate
- Recipe: Drug trafficking gains laundered through real estate purchases, inflating prices
- Also happening in Missouri, Kansas, Dubai
US Drug Crisis
- Drug overdose #1 cause of violence in US
- Illicit drug users spent $150 billion on drugs
- Over 9.2 million shipping containers with narcotics reach US coasts annually
- Sources: Mexico, Myanmar, Afghanistan
- Current enforcement methods ineffective:
- Random patrolling
- Opportunistic human intelligence
- Outdated policing techniques
Criminal Technology Adoption
- Criminals using advanced technology while law enforcement uses basic methods
- Spanish Mafia: Autonomous submarines for drug trafficking logistics
- Embargo violators: Satellite phones for communication
- Pirates: Satellite systems to detect, chase, capture vessels for ransom
Hava AI System Development
- Created with business partner
- Uses 650,000 keyword combinations in 23 most spoken languages
- Functions:
- Searches crimes online in all available databases
- Analyzes press releases, Interpol notices
- Identifies involved parties, locations, ship routes, identities
- Maps criminal networks
- Development process:
- Consulted ethicists to reduce bias
- Spoke with law enforcement about gaps
- Interviewed criminals to understand evasion methods
Vessel of Concern Approach
- Identifies dangerous/potentially criminal vessels entering waters
- Detects shady companies wanting to do business
- Provides ability to see high-risk ships on map even when appearing innocent
- Similar to profiling known persons of interest or tracking repeat sex offenders
Case Studies & Results
Madagascar 2022
- Refused fishing licenses to Chinese vessels with criminal records
Spanish Companies
- Multiple companies in "murky waters" after illegal fishing and labor offenses uncovered in Senegal
Fu Yuan 831 Vessel Case
- Caught in Indonesia on multiple charges
- Flying multiple flags
- Owner: Chinese elected Congressman
- Complex ownership network to hide involvement:
- Shell company in Timor Leste
- Partly owned by Chinese state-owned company
- Convoluted structure common for crime concealment
Maritime Crime Trends
- Maritime crimes tripled in past 10 years
- Common failed solutions:
- "Stop eating fish to save ocean"
- Send more boats at sea
- Close borders, build walls
- These approaches don't work for maritime drug trafficking
Technology-Human Collaboration Necessity
- Technology powerful but lacks nuance, understanding, knowledge
- Human networks stored in individual minds essential
- Spanish companies investigation example:
- Technology showed vessel locations, speeds, identities
- Still needed fishermen interviews for political context
- Thailand findings: 25% of fishing laborers worked 17-24 hours/day nonstop
- When catching illegal fishing vessels, must remember crew may be human trafficking victims
- Need human approach to complement technology
- Target right criminals - what we see doesn't tell whole story
Conclusion
- Human creativity and technological progress must go hand in hand
- Technology helps tackle crime, save time/money, share knowledge, build bridges
- Only works when people fully involved in process
- 15 years working with agencies, fishermen, criminals confirms this approach
