If you want to pay by credit card
This paper does not concern the utilization of artificial intelligence tools by attorneys. Their proper employment, the verification of outputs, and the professional liability arising therefrom constitute a separate matter. The subject matter of this paper pertains to the activities of "non-attorneys"—namely, entities lacking the legal authorization to provide paid or gratuitous legal counsel, opinions, or recommendations (specifically, artificial intelligence).
I would also note that, akin to the dispute between taxicab operators and Uber, I do not advocate for professional bodies attempting to obstruct technologies beneficial to society out of apprehension that such technology will "usurp our profession." Nevertheless, examining the developmental trajectory of artificial intelligence within the legal sphere over the past year, its prospective implementations, and the public reaction thereto, I am of the opinion that the most prudent course is for this evolution to proceed within defined parameters and over a measured timeframe. In this regard, a substantial burden devolves upon bar associations, which are tasked with shaping the nascent regulatory framework.
Beyond these considerations, however, the fundamental inquiry remains: "Are the activities of artificial intelligence (and its operating corporate entities, etc.) within the legal domain consistent with the Attorneys' Act?" Should such activities contravene the statute, they must either be enjoined, or the Attorneys' Act must be amended in accordance with public interest. The delicate balancing test between the legal profession and public interest—and how these competing weights are to be weighed—warrants rigorous scholarly debate. This issue, however, transcends the scope of the present inquiry.
The heading of Article 35 of the Attorneys' Act is titled "Services Reserved Exclusively for Attorneys." The relevant provision explicitly dictates: "The rendering of opinions on legislative affairs and legal matters... the conduct of judicial proceedings, and the drafting of all documents pertaining thereto, shall be restricted exclusively to attorneys registered on the roll of the bar." Conversely, paragraph 1 of Article 63 of the Attorneys' Act provides that: "Persons not inscribed on the bar roster... may not draft litigation pleadings pertaining to matters other than their own... nor exercise other prerogatives vested in attorneys."
Paragraph 2 of Article 63 categorizes this conduct as a regulatory offense (kabahat), subject to an administrative fine as its sanction. It is noteworthy that while the primary penalty for the unauthorized exercise of attorney-reserved prerogatives is an administrative fine, the forfeiture (confiscation and transfer of ownership to the state) of instrumentalities may also be triggered pursuant to Article 18 of the Misdemeanors Act.
Pursuant to Article 63 of the Attorneys' Act, bar associations apprised of unauthorized legal practice must report such circumstances to the Office of the Chief Public Prosecutor. Furthermore, under Article 76 of the same Act, the statutory mandate of bar associations encompasses the advancement of the profession and the safeguarding of its professional standing and prestige.
Even artificial intelligence tools made available pro bono directly furnish legal advice, draft legal memoranda, draw up contracts, and draft pleadings. Commercial counterparts engineered explicitly for legal applications are likewise commercially available and are not restricted solely to the use of licensed attorneys.
Under the Attorneys' Act, any individual citizen possesses the autonomy to draft and file their personal pleadings pro se. However, in my estimation, whether a pleading generated via artificial intelligence can genuinely be deemed "drafted" by the citizen remains highly questionable. In essence, the pleading is formulated by the AI software—that is, the operating corporate entity—leveraging the user's disclosures combined with a substantive legal knowledge base. The analysis applies equally to formal legal opinions.
Inquiry 1: What distinguishes this activity from the drafting of pleadings by traditional petition writers (arzuhalci) or the guidance and drafting services rendered by insurance claim consultancy firms? A potential counter-argument posits that "there is merely software involved, which the citizen independently utilizes."
Consider, then, the analog scenario of a traditional petition writer stationed outside a courthouse drafting petitions. After conducting an intake interview regarding the grievance, the writer drafts the pleading and hands it over; the client fills in certain particulars and personally submits the document to the front desk or court registry. Are such commercial acts compatible with the Attorneys' Act? Based on the historical enforcement posture of bar associations, presumably not.
Let us modify the hypothetical slightly. Suppose our petition writer develops an AI software and installs a coin-operated terminal outside the courthouse. Citizens access the terminal to have their pleadings generated, print them out, and file them at the court registry. Does the petition writer's enterprise remain compliant with Article 63 of the Attorneys' Act? (Substitute the corporate entities operating the AI for the petition writer in this hypothetical, and reevaluate your conclusion.)
Inquiry 2: Does the inclusion of a "not legal advice" disclaimer in the terms of service of artificial intelligence software validate its operations under the Attorneys' Act? When queried regarding a legal controversy, artificial intelligence renders definitive assertions—typically dictating "this is the law – that is the rule"—rather than qualifying statements such as "this is speculative – do not construe this as professional advice – consult counsel." Given that even practicing attorneys cannot guarantee judicial outcomes, would the appending of a standard caveat beneath every response suffice to legalize the conduct?
Returning to our petition writer hypothetical: if this individual prefaced their services by declaring that the drafting is "non-advisory," would the resulting complaints and answers fall outside the scope of the unauthorized practice of law?
Consider a further variation: Suppose a layperson hosts a broadcast responding to legal queries, accompanied by a scrolling onscreen disclaimer reading "for informational purposes only." Would the rendering of legal counsel and live drafting of pleadings by a non-jurist withstand scrutiny under the Attorneys' Act?
Conclusion: I submit that these and parallel inquiries—alongside the overarching determination of whether AI activities constitute the unauthorized practice of law—must be authoritatively clarified through formal declaratory statements issued by bar associations.
Both from a regulatory standpoint and in terms of societal impact, there is no substantive distinction between the unauthorized dissemination of legal versus medical expertise.
When an AI model is solicited for a medical diagnosis, it typically responds along the following lines: "As an artificial intelligence assistant, I am not capable of providing medical diagnoses, evaluating pathology, or issuing prescriptions. This finding strictly necessitates professional medical evaluation."
Conversely, when prompted to draft a legal pleading, the same system produces a comprehensive document and, upon request, even computes litigation success probabilities. Similarly, when instructed to prepare a legal memorandum, it delivers a detailed juristic opinion. This dichotomy is profoundly disconcerting.
Civil litigation has been instituted against OpenAI in the United States concerning the dispensation of legal counsel via ChatGPT and its utilization in judicial proceedings. Concurrently, legislative bills have been introduced in New York to curb analogous infractions.[1] Consequently, the legal community faces a substantive exigency demanding rigorous institutional response.
Atty. Alp Öztekin
