DarkNPS: The Legal Implications of AI Within Drug Enforcement

By Abraham Hilu '25

---

Drug overdose casualties not only affect the victim but are also psychologically traumatic for the victim's friends and family members. Recently, a close friend of mine overdosed on a drug that was laced with fentanyl. I did not hear from him for days, yet I had a suspicion that something dangerous had happened to him. I received a text from another close friend that he had been hospitalized shortly after we hung out with him. Thankfully, my friend had survived this intense episode, but for a split moment, I had imagined a reality where I had lost him.

Much like most American victims of drug overdoses with illegal substances, my friend was a victim of a bigger issue that policymakers are still attempting to tackle. More than one million people have died since 1999 from a drug overdose. (1) Information is restricted to the individual to create proper decisions on actions they decide to make. With technologies on the rise in solving public health issues, one specifically named DarkNPS is an algorithm that detects possible new drug compounds that are created from illicit labs. The lead researcher of DarkNPS, David Wishart, asserts that, as soon as a new drug is introduced to the black market, DarkNPS could “potentially identify the compounds in a matter of minutes.” (2) Wishart believes that with the application of this emerging technology used in government agencies such as law enforcement, this method will be a great approach to solving the issue of drug-related deaths in the United States.

DarkNPS is a technology with a run command to detect possible variations of drugs. The reason behind the creation of this technology was to respond to drug makers in illicit labs who were able to create a different variation of a common drug. For example, “instead of PCP, they might have 3-MeO-PCE.” (3) Once DarkNPS is able to detect the new drug variation, government agencies will have the ability to create legislation that will inhibit the new drug from becoming rampant among citizens. This is a “fight fire with fire” approach; DarkNPS is using data to formulate new drug compounds much like drug makers in illicit labs. DarkNPS generates theoretical drug variations through data consumption of chemical compounds and potential combinations through machine-learning algorithms. This would then prepare the Local, State, and federal governments to create a legislative response.

DarkNPS intends to work alongside law enforcement and state officials to be the main ingredient in solving the war against drugs. This could have negative implications. These agencies and the creators of DarkNPS should fully acknowledge the notorious history of racial profiling of minorities as it relates to drug criminalization. In addition, they should fully consider this history of knowledge and implement reform within their strategy. This concern presents a deep connection to another issue at hand, which is the mass pooling of minorities who are legally prosecuted for drug felonies. A report published by the U.S. The Department of Justice shows that “among persons ages 20 to 24, black males were imprisoned at about 7 times that of white males.” (4) Not only are minorities unduly prosecuted, but they are also subject to cruel prison conditions.

Before 1995, federal courts would remediate unconstitutional prison conditions with structural inductions requiring states to comply with constitutional requirements. This would lead to a reform act titled, “Prison Litigation Reform Act of 1995” (henceforth referred to as PLRA). Brown v. Plata had cemented the Prison Litigation Reform Act of 1995, reinforcing the assertion that California prisons were violating the Eighth Amendment by overcrowding their prison systems. (5) For further context, California prison systems (referred to as CPS) from 2001-2011 had exceeded normal capacity by over 200%. In Brown v. Plata, a group of prisoners in California with serious medical conditions filed a complaint that there was a lack of medical care for treatment. They had noted that, due to overcrowding, medical treatment was practically impossible, requesting a reduction of prisoners in CPS under the clause of the PLRA. (6) This group had requested remediation under a 3-court judge circuit. The court had ordered to reduce the number of prisoners from 200% to 137% of its carrying capacity, releasing around 46,000 prisoners. (7) California appealed to the Supreme Court on the basis that federal judges do not have authority under the Constitution to run state penal systems, and they should be under the obligation of the State to protect those acts. The Supreme Court reaffirmed the 3-court judge circuit of California and stated that the outcome of prison overcrowding violated the Eighth Amendment under cruel and unusual punishment.

In March 2010, President Obama passed the Fair Sentencing Act (S. 1789), which would decrease the disparity ratio sentencing of crack and cocaine offenders from 100:1 racial disparity down to 18:1. (8) While there remains great racial disparity in cocaine sentencing against African Americans, this has become a crucial step in recognizing the unjustifiable discrimination and harsh sentencing of minorities in incarcerated systems. The history of prison reform is an important subject to consider when creating future drug reform since it directly correlates with prison systems.

As we fast forward to today, the implementation of AI has been a key factor in developing new strategies for tackling systemic issues facing the United States. Ruha Benjamin, a Sociology and African-American professor at Princeton University speaks on how detection algorithms have been used against minorities in her book, “Race After Technology''. In the chapter, Technological Benevolence Do Fixes Fix Us? Benjamin develops a theme in her book noting that “the more things change, the more things they stay the same.” (9) She explains the introduction of AI use for law enforcement from “mass incarceration” methods into a new form of oppressive practice known as “mass monitorization”. Since overcrowding methods had become unconstitutional as previously mentioned, prison systems implement Electronic Monitoring (EM) on those who are prosecuted for crimes. Moreover, EM is another use for state authority to control prosecuted minorities. Executive director of the Center for Media Justice, Malkia Cyril, posited “There is increasing evidence that the algorithmic formulas used to make these decisions [about who should be assigned e-monitoring] contain deep-seated racial bias.” (10)

The connection that DarkNPS has with these technology applications is law enforcement misapplying these tools and using them to reinforce their racial bias in drug crime. Dark NPS is capable of predicting the next illegal drug on the market, and law enforcement is capable of enacting drug operations to find these new drugs. Both of these aspects intersect, one tool will bolster the other highly driven agenda. A Washington Post article that researched DEA operations in low-income neighborhoods targeted for drug trafficking discovered that in “the last decade, none of the 179 defendants arrested in DEA reverse-sting cases in the Southern District of New York were white.” (11) More than often, these neighborhoods are composed of ethnic minorities. This gives the appearance that these neighborhoods are racially targeted. This ultimately worries the courts since government agencies are misusing their authority under constitutional provisions. Until these conflicts have become resolved with proper monitoring, providing powerful tools such as DarkNPS will further racial discrimination in dense minority neighborhoods such as areas in NYC.

With the combination of the risk potential for illicit labs to use emerging technology in their favor and the history of using racial prejudice to incarcerate minority groups, DarkNPS users should highly consider these factors when deciding the correct procedure for solving the illegal drug epidemic. The focus should be centered on the drug that is currently damaging American lives. Algorithms have a predictive nature in what "could be" instead of what should have never been a possibility in the first place. Not only do algorithms gear law enforcement to unjustly confirm a suspect based on racial prejudice, but it also distracts more relevant causes for illegal drug trafficking. Why should we risk the freedoms of individuals and create more harmful information on what new drugs can be created? Using these criticisms, we can deduce that using DarkNPS and other EM technologies can cause more potential harm than potential good in solving the increase of drug-related deaths in the United States.

We should not compromise minorities' freedoms to solve a systemic problem that does not showcase clear evidence of solving the war on drugs. DarkNPS can potentially be used to racially discriminate against innocent minorities who do not deserve ongoing accusations and investigations. Let us not detract and fuel more fires. I do not see positive results in using DarkNPS as a central method in solving drug-related issues without heavily considering the legal implications. Let us have consideration for citizen's freedom when solving important issues such as drug overdose outbreaks. By acknowledging these concerns we can create methods that both decrease drug overdose outbreaks while also reducing unjust criminalization of minority victims.

Endnotes

1. CDC. “Drug Overdose Deaths.” Centers for Disease Control and Prevention, August 22, 2023. https://www.cdc.gov/drugoverdose/deaths/index.html#:~:text=Opioids%E2%80%94mainly%20synthetic%20opioids%20(other,of%20all%20drug%20overdose%20deaths).

2. Johnson/Undark, Doug. “This Algorithm Knows about Recreational Drugs That Don’t Even Exist Yet.” Popular Science, December 29, 2021. https://www.popsci.com/health/algorithm-drug-overdoses/.

3. Ibid.

4. E. Ann Carson, Ph.D., and William J. Sabol, Ph.D., BJS Statisticians “Prisoners in 2011.” Home, December 2011. https://bjs.ojp.gov/content/pub/ascii/p11.txt.

5. Brown v. Plata, 563 U.S. 493, 131 S. Ct. 1910, 179 L. Ed. 2d 969, 2011 U.S. LEXIS 4012, 79 U.S.L.W. 4320, 22 Fla. L. Weekly Fed. S 995 (Supreme Court of the United States May 23, 2011, Decided). https://advance-lexis-com.lp.hscl.ufl.edu/api/document?collection=cases&id=urn:contentItem:52XN-CJP1-F04K-F3NP-00000-00&context=1516831.

6. Ibid.

7. Ibid.

8. ACLU. “President Obama Poised to Sign Bill Reducing Cocaine Sentencing Disparity after House Passage.” American Civil Liberties Union, July 27, 2010. https://www.aclu.org/press-releases/president-obama-poised-sign-bill-reducing-cocaine-sentencing-disparity-after-house.

9. Benjamin, Ruha. “4 Technological Benevolence Do Fixes Fix Us?” Essay. In Race After Technology, 140–45. Cambridge, UK: Polity, 2019.

10. Ibid.

11. Jacobs, Shayna. “10 Years. 179 Arrests. No White Defendants. DEA Tactics Face Scrutiny in New York. - The Washington Post.” 10 years. 179 arrests. No white defendants. DEA tactics face scrutiny in New York., December 14, 2019. https://www.washingtonpost.com/national-security/10-years-179-arrests-no-white-defendants-dea-tactics-face-scrutiny-in-new-york/2019/12/14/f6462242-12ce-11ea-bf62-eadd5d11f559_story.html.