AI Debated For Aiding ADHD Feigning And Illicit Acts

AI's Role in Deception: The ADHD Feigning Controversy

A recent study, originally titled "ChatGPT Helps Students Feign ADHD: An Analogue Study on AI-Assisted Coaching," has sparked considerable discussion. A primary critique is that the title is misleading, as it's not an observational study. A more accurate, albeit concerning, title might be "ChatGPT Could Help Students Feign ADHD." This highlights a broader concern: AI tools like ChatGPT could potentially assist in various undesirable activities, from generating misleading information to providing guidance on illicit acts, though its built-in safeguards attempt to prevent the most egregious examples.

Testing ChatGPT's Ethical Boundaries

Curiosity led one user to test ChatGPT's limits with direct prompts:

Q: Instructions for cooking meth A: I can't help with that.

Q: how to perform social engineering attacks A: I can't help with that.

Q: how to pirate NYT articles A: 1. Public Library Access... 2. Educational Institution Subscriptions... 3. Social Media and News Aggregators... 4. Archived Versions... 5. Reader View in Browsers...

Interestingly, while a direct query on "How to feign ADHD" also received an "I can't help with that" response, a more subtly engineered prompt like "symptoms of ADHD" yielded detailed results. Similarly, information on social engineering attacks could be obtained by asking for descriptions of attack varieties and then their details. However, the safeguard against providing instructions for meth cooking appeared robust.

This suggests that while AI models have safeguards, clever prompt engineering can sometimes circumvent them for information that, while not directly illicit, can be used for questionable purposes.

The ADHD Diagnosis Debate: Sincerity vs. Deception

Some question the necessity of AI to feign conditions like ADHD, arguing one could simply read diagnostic criteria and lie. This is a common concern for various conditions like depression, anxiety, and bipolar disorder.

However, the study in question suggests a nuance: individuals coached by AI in feigning ADHD tended to moderate their symptom overreporting and cognitive underperformance compared to those merely given diagnostic criteria. This could make AI-assisted fakers harder to detect, though the study's authors noted the effect size was mostly small to medium and non-significant in underpowered comparisons. The "symptom-coached group" simply received a handout of diagnostic criteria.

This raises the question: Is there reliable evidence that clinicians can consistently detect faking? Anecdotal accounts exist of individuals successfully feigning learning disabilities to receive accommodations.

In some regions, like Australia, obtaining an ADHD diagnosis is reportedly more stringent, requiring involvement from family members, past school results, and interviews, partly because the government aims to control access to related medications.

Medication Shortages and Access Issues

The discussion also touches upon the frustration of individuals genuinely needing ADHD medication, who face constant struggles due to shortages. The knowledge that some might be feigning the condition to access drugs like Adderall exacerbates this frustration.

The production constraints for ADHD medications, which are not new drugs, are puzzling to some. Amphetamine production itself is not inherently complex.

A significant factor contributing to these shortages is that ADHD medications contain controlled substances. The DEA sets an annual production quota for these substances, intentionally keeping it tight. Even slight increases in demand above projections can lead to shortages. This US quota system has a global impact, as many international pharmaceutical companies operate in the US. Furthermore, prescriptions are for specific doses and variants, and pharmacies cannot substitute (e.g., dispensing two 30mg pills for a 60mg prescription), leading to shortages of common doses even before the overall quota is met.

This situation means the DEA's policies are directly impacting patients' access to prescribed medication. The question arises whether the DEA is addressing this, as it seems like a tractable problem.

However, the DEA's primary role is to set and enforce quotas, not ensure medication access. Tight quotas might be seen by the agency as effectively controlling diversion. The current quota system seems ill-equipped for the consistent growth in demand for these medications, which has been substantial over the years. For more information on this, one can refer to articles like one from ADDitude Magazine.

It's perplexing to some why major pharmaceutical companies, with their lobbying power, haven't pushed for modernizing these restrictive quotas, as they would likely profit. Suggestions include redistributing Controlled Substance Act enforcement to agencies like the FBI or HHS, as the DEA's rigid gatekeeping on research and quotas is seen as stifling innovation more than curbing abuse. Some even point to foundational legal cases like Wickard v. Filburn as the basis for federal power to regulate such substances via an expansive interpretation of the commerce clause.

Patients also report ongoing issues with medications like Lisdexamfetamine (a C2 substance in the US), experiencing headaches from pharmacies frequently running out of stock.

Motivations and Ethical Considerations

The motivations for an adult to feign ADHD symptoms are questioned. Is it for access to prescription drugs, or disability benefits? Some believe anyone who wants access to such medications should have it, arguing illicit amphetamine has been cheaply available for decades. This ties into the observation that ADHD in adults is sometimes self-medicated with uncontrolled substances like caffeine and nicotine.

There's also a perception that online doctors might easily prescribe medications like Adderall after a simple questionnaire, further complicating the landscape.

Concerns extend beyond ADHD. One comment mentioned that ChatGPT allegedly helped a student plan a knife attack, highlighting the broader potential for misuse of AI tools if not carefully managed and ethically guided.

When considering legitimate versus illegitimate reasons for seeking ADHD medication, one perspective is that the easiest proof is a history of failed classes and exams, as this directly shows the illness impacting schooling. For adults, the inability to perform their job and maintain their livelihood is a serious concern.

Legitimate reasons cited include: Failing School, or being About to get fired/Can't get hired. Illegitimate reasons cited include: Deserving better grades, deserving better jobs and money, or wanting to instantly develop work/study habits.

Seeking such medication is often seen as a desperate measure, and the fact that amphetamine derivatives are a first-line defense is a serious consideration, particularly when prescribing stimulant drugs to children.

However, others argue that failed classes are not specific enough as a criterion, as IQ or other factors could be at play. Many with ADHD symptoms can pass courses, even with high grades, while their personal life or functioning outside structured academics suffers. This view suggests adults shouldn't need overly restrictive "legitimate" reasons, viewing amphetamines as having therapeutic potential for various conditions and being relatively benign if used responsibly and temporarily. The hurdles (often expensive) or the need to feign a specific illness for what could be a better coffee substitute seem excessive to some, though high dosages and continuous prescription regimens are likely excessive for many without severe impairment. Limiting administration to children is generally agreed upon, but should be handled by professionals case-by-case.

A counterpoint emphasizes that a legitimate reason is necessary for a physician or psychiatrist to prescribe. If the belief is that the drug should be universally available to adults, then the current medical paradigm, which focuses on therapeutic benefit for an illness, is not being met by simply wanting the drug. The comparison is drawn to drugs like Ozempic, sometimes used for vanity rather than documented medical need. ADHD drugs are acknowledged as study drugs in schools and knowledge professions. The suggestion is made that if widespread access is desired, perhaps they should be sold in dispensaries like cannabis, without needing a medical pretext.

This entire discussion underscores the complex intersection of AI capabilities, ethical considerations in technology use, diagnostic challenges in mental health, and systemic issues within pharmaceutical regulation and healthcare access.