Artificial intelligence (AI) is having quite the moment. Think-pieces around the country are predicting its effect from employment, traffic, financial markets, security, writing, education and even crime.
To be sure, AI’s potential to change our lives is wrapped up in a lot of fear around the technology replacing or controlling us. Blockbuster films and franchises such as Minority Report, the Matrix Trilogy, Ghost in the Shell and more put our worst nightmares about technological control on the silver screen.
Now, science fact is catching up with science fiction, as AI has already made headway in pre-crime initiatives, known as predictive policing.
This is the use of AI to potentially make an arrest — or at least flag a suspect for questioning — before a crime is even committed. This represents a new form of policing which has incredible potential for improving our lives, as well as incredible potential for abuse.
Artificial Intelligence and Pre-Crime
There are multiple methods of how this artificial intelligence works, but in a nutshell, it takes a list of known criminals, their movements and general location, and it triangulates that against maps of areas where crimes are most likely to occur. It then creates a list of individuals for intervention by law enforcement.
Similar software can also be used by judges when considering sentences for criminals and by parole boards when considering the eligibility of a convicted individual for release.
A slightly different and less invasive philosophy of pre-crime policing is the identification of specific areas and times in those areas where crimes are likely to be committed. Police then increase their visible presence in those areas to deter crime from occurring.
However, it’s not only law enforcement that takes advantage of AI’s pre-crime capabilities. Retailers are also partnering with AI providers to spot shoplifters prior to the actual theft of the product. In fact, Japanese startup Vaak uses AI to identify potential shoplifters using body language and other physical movements.
AI on the high seas
AI aids in identifying other nontraditional forms of crime as well.
With over 70% of the earth’s surface covered by ocean, ships, with their long range and massive cargo capacity, continue to play an important role in the lives of virtually everyone on the planet. Unfortunately, due to nefarious corporations and nation states, shipping in the 21st century still has ties to its pirate roots:
Fishing
- Illegal fishing in protected areas
- Overfishing
- Enslaved crews on fishing boats
Sanction Avoidance
- Exportation of oil when prohibited by international sanctions
- Importation of goods when prohibited by international sanctions
Smuggling
- Counterfeit and prohibited goods
- Drugs
- Exotic flora and fauna
- Humans
- Weapons
To combat this, AI is being used in very innovative ways.
For example, Israeli startup Windward utilizes millions of data points to train its AI, which gives each ship a unique fingerprint. This allows for greater tracking despite efforts to obfuscate the identity, location and cargo of individual ships.
Additionally, Global Fishing Watch (GFW) has tracking platforms that harness the predictive and unique characteristic-identifying power of AI to police the ocean the way you would a city. GFW tracks “fish laundering” where smaller ships offload their catches to larger ships, allowing them to stay out longer. This also makes it easier for them to hide overfishing and fishing in protected areas.
By monitoring actions such as the frequency in which a ship launches its nets and total days at sea, investigators can more accurately determine whether the ships are using an enslaved crew as well.
AI on the highways
With an estimated 287 million registered vehicles on the road in the US for 2020, it is clear that cars and trucks represent a treasure trove of potential data points begging to be analyzed.
It would also stand to reason that with that many cars and trucks on the road, hundreds of thousands, possibly even millions of vehicles are used to commit crimes annually. With that in mind, private companies are lining up to provide AI-powered solutions to law enforcement agencies (and even neighborhoods) to help solve crimes aid in pre-crime initiatives.
For instance, Hazen.ai is an AI-powered software that utilizes existing traffic cameras to identify everything from driving without using a seatbelt to texting while driving to a host of other potential events.
Additionally, companies such as Flock utilize proprietary cameras and software to fingerprint every car that passes within 75 feet of their installed cameras. The company states that just the presence of the cameras lowers the crime rate in the immediate vicinity. Flock not only markets their product to law enforcement but also to neighborhood associations.
Artificial Intelligence and Finance
Approximately 20% of finance firms have implemented artificial intelligence to reduce false positives in flagged financial transactions and fight fraud throughout their systems.
Money laundering is a crime with a massive international scope. Criminals have created sophisticated networks to move their money around in attempts to make it difficult to track. AI can cut down false alerts while drastically speeding up the investigation to thwart criminals and terrorists attempting to launder billions of dollars.
In personal finance, AI has been similarly effective; Artificial intelligence is powering advanced anti-fraud credit card technology that has progressed to the point where, despite millions of credit card numbers being sold on the dark web daily, actual successful fraud is relatively low.
Ethics
When it comes to using artificial intelligence in several of the examples mentioned above, there are clear benefits. However, the elephant in the room is ethics.
Americans are accustomed to a certain amount of privacy that erodes more and more by the day. Yes, some of it is performed willingly. A majority of adults carry smartphones and check in on various social media platforms everywhere they go, and they also enable location services to make it easier to find restaurants and other businesses.
However, the difference is choice – those forms of tracking can be turned off. Facial recognition, automobile tag readers and other AI-powered systems do not request people’s permission to log their activities.
Furthermore, pre-crime systems have been shown to be biased against minorities and those who have been convicted of a crime but have successfully completed their sentences. Some AI implementations of facial recognition have a eugenics-inspired feel to them, which further concerns ethicists.
One working solution: Fraud detection
AI can be implemented both ethically and in such a way that respects privacy by only using meta-data.
Fraud.net leverages pre-crime functionality in a number of ways, including account setups with information that is known to have been stolen to identify potential fraudsters before they buy. It also looks at the behavior of seemingly legitimate customers who may be exhibiting anomalous behavior as an early warning of fraud prior to a transaction.
To learn more about how you can detect pre-crime using Fraud.net’s sophisticated technology, download our FREE EBOOK.