Category: Law Enforcement

How Online Data & AI Help Fight Human Trafficking

Human trafficking is rampant in our country, but not often mentioned. It happens in the shadows, on the dark web, is hard to track, and tough to talk about.

The crime of human trafficking—defined by the U.S. Department of Homeland Security as modern-day slavery that “involves the use of force, fraud, or coercion to obtain some type of labor or commercial sex act—is notoriously difficult to prosecute. Victims, who are mostly women and children, often lack legal documentation and many fear reprisals if they go to authorities. Therefore perpetrators go to great lengths to conceal their behavior by laundering money and keeping their operations quiet. 

Recently, law enforcement agencies and organizations that help victims of human trafficking have begun using artificial intelligence tools to overcome this lack of visibility. By sorting through data and recognizing patterns faster than any human could, AI tools are helping activists and investigators crack down on buyers of sex, identify trafficking victims and follow illicit money trails. 


“Imagine the techniques that Google and Facebook are using to make a profit—understanding people, the way they connect, what their interests are, what they might buy or the activities they engage in,” says Dan Lopresti, professor and chair of the department of science and engineering at Lehigh University’s P.C. Rossin College of Engineering and Applied Science. “We can apply those same techniques—data mining, text mining, what’s called graph mining—AI that’s being used for legitimate and really profitable purposes, to track these illicit behaviors.”

Lopresti is involved with the Regional Intelligence and Investigation Center (RIIC) in Allentown, PA. Launched in 2013, the RIIC has “revolutionized” the way area police departments “analyze and share collected data to solve crimes,” according to the office of Lehigh County District Attorney James Martin.

Lopresti, who is an expert in document analysis and pattern recognition, is working with RIIC Director Julia Kocis, prosecutors, law enforcement officials and other Rossin College computer science and engineering faculty members to help overcome the challenges of turning vast amounts of data, primarily from police incident reports, into something useable, despite limited resources.

“If an expert sits down and reads enough of these, he or she will find a common thread—this person is related to this place, which is related to this activity, which is related to this other person,” he says. “The trouble is, they’ve got millions of these reports and just don’t have enough time to read through them. We’re developing natural language techniques, text mining and data mining techniques that are oriented to processing lots of data to identify patterns of behavior that would reflect illegal activities related to human trafficking.”

Because human trafficking is a $150 billion-per-year criminal industry, many companies are joining the fight and helping out. IBM worked with the STOP THE TRAFFIK (STT) coalition to develop a new cloud-hosted data hub that allows financial institutions to run AI and machine learning tools against their data sets to detect “specific human trafficking terms and incidents.” AI also allows the data hub to take in open-source data — including thousands of news feeds each day — to help analysts more easily identify the characteristics of human trafficking incidents. 

Josh McAfee

Entrepreneur Josh McAfee, a former law enforcement officer and loss-prevention executive, is the founder and CEO of McAfee Institute, a Peoria, Illinois-based business that trains corporations and law enforcement agencies to battle cybercrime and fraud. But what brings him the most satisfaction is locating victims of human trafficking.

Since 2011, his company has helped bring back about 1,900 human-trafficking victims, 1,000 of them children.

He now teaches a course on human trafficking to an average of 150 to 200 people a month. “We’re developing leaders in the classroom, so they can go out and utilize our techniques within their organizations. It multiplies our success and enables us to make a huge impact,” says McAfee. “Growing up, I always had a passion for law enforcement. I wanted to get out there and fight crime, get the bad guys–you know, help people. That’s why I became a law enforcement officer right out of college. To be able to do this kind of work in my business is everything to me. It’s what I was born to do.”

How Law Enforcement Uses The Dark Web

Have you ever had your identity stolen? If you haven’t, consider yourself lucky. On a daily basis we hear reports of someone stealing a credit card, hijacking bank accounts, or creating false online profiles. One thing is clear: Cyber criminals are constantly trying to steal our personal information…and this information often ends up on a place called the Dark Web.

So what exactly is the Dark Web?

The dark web is a portion of the internet that can only be accessed by using certain browsers and software.

NBC10 Boston Investigators sat down with Andrei Barysevich, the director of advanced collection for Recorded Future, a Somerville-based cyber intelligence company. “You can pretty much find anything,” Barysevich said. “Stolen identities, credit card numbers, compromised data or weapons and drugs.”

In the past two years, Barysevich has gone from a team of one to a team of several dozen cyber intelligence analysts, combing through more than 2 million Dark Web sources per week. His employees, typically fluent in several foreign languages, act like “flies on the wall” in Dark Web online forums, Barysevich said, attempting to gather information about what’s being bought and sold.

On one disturbing site, Barysevich showed NBC10 Boston Investigator Ryan Kath how easy it is to buy the Social Security number of almost anyone in the United States. Using one of Recorded Future’s accounts to pay the $3 charge, Kath plugged in his name. After a search that only took a few seconds, Kath’s personal information appeared on the screen. Barysevich said everyone should assume their information was at one point stolen and is available on the Dark Web.

Christopher Ahlberg, CEO of Recorded Future said, “Cyber security has grown incredibly in the past few years. It’s the idea of being able to catch cyber threats before they hit you. To do that, you need to infiltrate the places that bad guys hang out.” When valuable information is uncovered, Recorded Future shares the details with the pertinent parties, whether it’s a government agency, financial institution or law enforcement. One notable example was when the company spotted a hacker selling sensitive documents about military drones.

Mark Turnage is another investigator who is familiar with the dark web. Turnage’s company, DarkOwl, helps law enforcement and cyber security firms monitor the criminals who lurk in the dark web. “The dark net is appealing to criminals because it completely anonymizes their presence,” Turnage explained. There are no IP addresses. There is no way to trace the person on a keyboard to a physical location. All law enforcement can do is wait for criminals to slip up.

Luckily, that happened in the case against Danny McLaughlin. The Colorado man is accused of attempting to hold a 13-year-old girl captive for sex and trying to hire a hitman to kill his wife. The criminal complaint filed against McLaughlin says his crimes started on the dark web on a site for people who enjoy torture and killing.

McLaughlin was only identified and caught when investigators say he agreed to meet at a Centennial hotel with the person who had agreed to murder his wife. That person was really an undercover detective.

“Thank goodness he made that mistake and was arrested. Had he not gone to that hotel room, it would have been near impossible for them to figure out who this person was,” said Turnage.

Recently, International police group Interpol arrested nine people in Thailand, Australia, and the U.S. and 50 children had been rescued after investigators took down an online pedophilia ring. Police in nearly 60 countries combined efforts in this Interpol operation launched two years ago into a hidden “dark web” site with 63,000 users worldwide. Fifty children were rescued following the arrests.

Interpol said its Operation Blackwrist began after it found material that was traced back to a subscription-based site on the dark web, where people can use encrypted software to hide behind layers of secrecy. Interpol enlisted help from national agencies worldwide, with the US Homeland Security Investigations (HSI) department eventually tracking the site’s IP address, where new photos and videos were posted weekly.

The first arrests came in early 2018, when the site’s main administrator, Montri Salangam, was detained in Thailand, and another administrator, Ruecha Tokputza, was captured in Australia. Salangam, whose victims included one of his nephews, was sentenced in June last year to 146 years in prison by Thai courts. Interpol said children were lured to Salangam’s home with the promise of food, internet access and soccer games.  

One of his associates, a pre-school teacher, got 36 years. Tokputza was handed a 40-year prison term at his trial in Australia last Friday, the longest ever for child sex offences in the country. The Australian Associated Press reported that Tokputza, 31, pleaded guilty to 50 counts of abuse of 11 babies and children — one just 15 months old — between 2011 and 2018. The HSI’s regional attache in Bangkok, Eric McLoughlin, said in the statement that “numerous arrests” had been made in the US. Some held “positions of public trust,” he said, and one individual was abusing his two-year-old stepbrother.

“Operation Blackwrist sends a clear message to those abusing children, producing child sexual exploitation material and sharing the images online: We see you, and you will be brought to justice,” Interpol’s Secretary General Juergen Stock said.

What Can You Do to Avoid the Dark Web?

To reduce your risk of being hacked or having information stolen, Barysevich offers these tips:

• Freeze your credit report, something that can be done for free

• Activate text and email alerts for activity on your bank accounts

• Question why you need to provide a Social Security number or copy of your driver’s license when you visit the doctor, dentist, or other professional office

• Don’t use the same password for multiple online accounts

While some consumers might want to throw up their hands in frustration, Ahlberg — the Recorded Future CEO — said not to give up the fight. The goal is to decrease the odds that you’ll be cyber thieves’ next target.

But That’s Not Our Guy – Why Social Media & Online Research Can Be Frustrating

DIY platforms and other data gathering programs are great, but before you consider them to be inaccurate or not helpful, keep in mind that they are only as good as the person using them for research.

For many professions, including private investigators, law firms, and insurance companies, social media and online research on individuals has opened a new venue for assisting with investigations into court cases, insurance fraud, and other situations.

While it can be done, it can also be time consuming and frustrating. The more common the name, and the less information that is being used to search for a person’s online activity, the longer it can take. One roadblock is information that does not actually belong to the person you’re looking at – their information is getting mixed up with someone else with the same name.

In talking with folks in the industry, this is a pain point that not only makes it more difficult to search for people online, but it also gives the impression that the quality of such searches is not great. This can make investigators leery and question the benefit of conducting such research.

It’s important to remember a few key points when considering options for conducting social and online research:

1. Software is good, but will never be perfect. Whether it’s a DIY online search platform or TLO/Tracers/IDI report, there is a chance that information will be provided that does not in fact belong to the individual you’re interested in, whether it’s a phone number, email address, or social site. The opportunity for false positives is there no matter what platform is being used.

It’s important to utilize multiple software platforms – what one will pick up, another will miss, and what one provides as a “false positive” another will not pick up at all. Each platform works off its own algorithm, or process for producing search results based on what information is provided about an individual.

If an investigator is using such programs to conduct a search, they can be very valuable. However, it is important to keep in mind that not all information provided will be accurate; these are meant to be used as a starting point in an online investigation – manual research & identification will be needed to confirm whether a piece of information really belongs to the person you’re looking at. Typically, if it’s not readily clear from the initial search, a good rule of thumb is to find three pieces of confirmation to ensure it’s the person’s social account or site.

Here’s an example: if you find a Facebook page that you think belongs to the individual, but perhaps it’s not really clear just from looking at the user name or “About” section, take a look at their friends list, places they’ve checked in, “About section” to see if birthdays or employers are listed, etc. If you can identify the location of the individual whose Facebook page you’re looking at, or maybe confirm that some of their Facebook friends are relatives that correspond with findings from a TLO report, then you can be more certain that it’s the right account.

2. What may seem to be inaccurate information can actually be a key to what you’re looking for. Sometimes people will see social reports and say something like, “That phone number for Joe is so old and hasn’t been used in years. This report isn’t accurate at all. I want his current phone number.” Valid point, but here’s something to keep in mind: Joe may have accounts, comments, or forum membership tied to these old numbers. While he’s not currently using the phone number, the accounts still live online and are easier to tie to Joe. Sometimes these old accounts are long forgotten, which means they’re not being scrubbed if they are involved in a situation where a lawyer tells them to “clean up their online presence.” This means that there could be valuable information to find based on what seems to be inaccurate information.

3. DIY programs are great, but they are called DIY for a reason. There are many companies who will allow you to purchase a subscription to an online search product to find content posted by or about an individual. As this type of research has evolved, the services have greatly improved. There is definitely a place for do it yourself search programs – but you have to do it yourself. No matter what the company claims, it will not be as simple as entering some information about an individual and being presented with all of the person’s online activity without any false positives or inaccurate information. It’s going to take you (the “yourself” part in DIY) to validate, investigate, and determine the validity of the results. Before deeming a DIY program as not useful, remember its actual purpose and that it is not meant as a be all end all service. And, if the company is promising that it is, you may want to reconsider using it.

4. There is no magic bullet – online research takes a lot of time. There’s no way around it. While we all wish a software platform would be created to give us instant and completely accurate results, this will likely never happen. Why? Things are changing all the time, whether it’s social media privacy laws, Google algorithm updates, or any number of things that can change in an instant. This is where online and social media research gets tricky and frustrating, leaving people to give up easily. Software can not be relied on as a standalone product – manual research is needed to confirm the validity of the information provided, and then take that information as a starting point and fleshing out what can be found through manual searching. When multiple platforms and other similar products are used simultaneously, the time spent can be greatly reduced. This is why it can be useful to turn to a full-service social media and online research service – they often have efficiencies in place to search quicker and provide more accurate results, which saves a lot of time for those needing to conduct investigations.

DIY platforms and other data gathering programs are great, but before you consider them to be inaccurate or not helpful, keep in mind that they are only as good as the person using them for research. Take them for what they are and realize that they will not be the magic bullet to quickly investigate an individual’s online activity.

The Government Turns to Social Media for Social Security Fraud

Playing outside with your grandchildren, casting a fishing line or running a marathon can seem like harmless, healthy fun. But physical activity could lead to denial of Social Security disability benefits if your activity shows up on Facebook or Instagram. Is it fair for the government to go through your social media? On the other hand, is it fair for you to apply for benefits if you can do these activities?

These are all questions the Social Security Administration is weighing, as Acting Social Security Commissioner Nancy A. Berryhill has told Congress in written explanations over the last week. Facebook and other social media feeds are already being reviewed if agency investigators suspect someone is fraudulently collecting benefits and they are looking for corroboration, the agency said in the documents it gave to lawmakers. But now the agency is “evaluating how social media could be used by disability adjudicators in assessing the consistency and supportability of evidence in a claimant’s case file,” Berryhill said in the submission to Congress.

The idea has drawn praise as well as criticism — praise because more attention to people’s social media could cut down on fraud, some say. Fraud and abuse do exist in the program, and it should be weeded out to protect taxpayers and legitimate claimants. 

The Pros

Mark Hinkle, acting press officer for the SSA, notes that the agency uses data analytics and predictive modeling to detect fraud, and has created new groups dedicated to detection and prevention. Asked to comment on plans for expanded use of social media to detect fraud, he confirmed that SSA investigative units already use social media, and that the agency has “studied strategies of other agencies and private entities to determine how social media might be used to evaluate disability applications.”

The Social Security Administration said that in December, 8.5 million people of all ages received a total of $10.5 billion in disability benefits, with an average monthly sum of $1,234. It said 68 percent of the private-sector workforce has no long-term disability insurance.

The risk of disability rises with age, and people are twice as likely to collect disability benefits at age 50 as at age 40 — and twice as likely at age 60 compared to age 50, according to the Center on Budget and Policy Priorities, a liberal Washington think tank.

“Social media can provide valuable evidence to support or deny individuals’ disability claims,” Rachel Greszler, a research fellow at the Heritage Foundation, a conservative Washington think tank, wrote last year. “For example, a disability claimant may say that she is unable to leave her home, while her social media pictures show her out and about regularly.”

In one case, a 57-year-old Louisiana man pleaded guilty last month to theft of government funds. He had received $2,177 a month in benefits — a total of $242,000 — while employed by companies that did demolition work and job site cleaning. He also operated heavy construction equipment. He told federal investigators that the companies had been registered in the names of family members, rather than his own name, “so y’all wouldn’t find out about it,” according to court records.

In its latest financial report, Social Security estimated that it made $3.4 billion in overpayments to disability insurance beneficiaries in 2017, in part because of their failure to report work activities. The program has been “riddled with problems, including fraud and abuse,” said Greszler. When people who can work collect benefits, she said, “it drains the system for those who truly cannot work and support themselves.”

The Cons

Advocates for people with disabilities say the use of social media in this way would be dangerous because photos posted there do not always provide reliable evidence of a person’s current condition. Someone may not want to upload a picture or video that shares how he or she deals with a disability on a daily basis.

Social media reviews could also delay the time it takes for applicants who are already out of work to be approved. “This proposal starts with the discriminatory assumption that people with disabilities do nothing socially in the community, or have lives, so that anything the person does on social media can be classified as some form of fraud,” said Eric Buehlmann, deputy executive director of public policy at the National Disability Rights Network, a nonprofit membership advocacy group. He told AARP that methods to detect fraud already exist. If the Social Security Administration “is interested in rooting out fraud,” Beuhlmann said, “spying on people’s social media accounts is not the way to do it.”

“It may be difficult to tell when a photograph was taken,” said Lisa D. Ekman, a lawyer who is the chairwoman of the Consortium for Citizens with Disabilities, a coalition of advocacy groups. “Just because someone posted a photograph of them golfing or going fishing in February of 2019 does not mean that the activity occurred in 2019.”

Program statistics do not support the allegation that SSDI is riddled with fraud and abuse. In the government’s fiscal-year 2018, the SSA’s Office of the Inspector General (OIG) reported about $98 million in recoveries, fines, settlements/judgments, and restitution as a result of Social Security fraud investigations. The OIG states that most the recovered funds were from recipients of SSDI and Supplemental Security Income (SSI), a means-tested welfare program for low-income seniors, blind and disabled people.

That sounds like big money. But in fiscal 2018, the SSA paid out $197 billion to beneficiaries of SSDI and SSI. And keep in mind that the recovered $98 million was for benefits paid out over several years, not just in 2018. SSA data shows that the rate of overpayments for all its programs was well under 1% of benefit payouts in each of the last three fiscal years – and not all improper payments are fraud.

More often, overpayments occur due to administrative delays at the SSA in making adjustments to benefit amounts due to errors and paperwork snafus. A federal government list of programs at highest risk for making improper payments compiled by the Office of Management and Budget does not even mention Social Security.

Greszler has readily acknowledged that fraud rates are low. “Outright fraud is actually a pretty small component of the program’s problems,” she said. “Most people perceive fraud as a big issue but what they might consider fraud – people receiving benefits when they have the ability to work – is often just abuse of the system by taking advantage of certain rules and structures that allow people who can perform some work to nevertheless receive benefits.”

What constitutes abuse of the rules? An example, she said, would be claiming SSDI and receiving unemployment benefits at the same time, or claiming based on the argument that a disability prevents a worker from performing certain types of jobs, Greszler and other SSDI critics often point to the rise of SSDI applications and award grants coincident with the rise of unemployment during the Great Recession as evidence of abuse.