Kim Kardashian Failed the Bar Exam. That's Not the Story.

Kim Kardashian announced in early December that she'd failed the California bar exam again. The internet had its predictable moment: memes, mockery, the usual pile-on. But Max Raskin, writing in the Washington Post, saw something else: proof that the bar exam is a cartel protection racket that should be dismantled entirely. 

His argument is seductive in its simplicity. Nearly half of California bar exam takers failed in July 2025. The exam tests arcane rules you can Google. AI already drafts contracts at "associate-level quality." Why force people through this gatekeeping exercise when the market can sort out competent lawyers from incompetent ones? 

Here's why: Raskin's argument confuses a flawed examination with the absence of need for any examination. And Ontario's experience over the past two decades offers a much more instructive lesson about what meaningful professional regulation looks like. 

The bar exam deserves criticism. Ontario is already fixing it. 

To be clear, traditional bar examinations have real problems. The Law Society of Ontario acknowledged as much in September 2025 when it proposed replacing bar exams entirely with a skills-based assessment program. LSO officials stated bluntly that current exams "are not designed to assess core skills required of lawyers, such as interviewing, legal research and writing, oral advocacy or negotiation." 

Candidates describe the experience as an "indexing exercise" rather than a test of understanding. Research from the Institute for Advancement of American Legal System confirms that "closed-book exams offer a poor measure of minimum competence" and "multiple choice questions bear little resemblance to the cognitive skills lawyers use." 

So, Ontario is moving towards scenario-based assessments, virtual law firm simulations, and evaluation by trained lawyer assessors throughout the licensing process. The proposal is currently under public consultation until January 31, 2026. Alberta, Saskatchewan, Manitoba, and Nova Scotia have already made similar transitions to the Practice Readiness Education Program. British Columbia announced in October 2025 it will follow suit, with PREP launching there in September 2026. 

This is reform, not surrender. The LSO isn't abandoning competency standards. It's making them more rigorous and more relevant. 

Raskin treats this distinction as irrelevant. His thesis is that any licensing requirement is just quantity control dressed up as quality control. But that's where his analysis breaks down. 

Professional judgment isn't something you can Google 

Raskin's central claim is that AI has made legal knowledge so accessible that licensing is obsolete. "AI systems already draft wills, nondisclosure agreements, term sheets, employment contracts and regulatory memos at associate-level quality," he writes. He dismisses concerns about AI hallucinations as "exceptions" and suggests lawyers are just embarrassed to admit they use these tools. 

The problem is that Canadian courts have been dealing with exactly these "exceptions" with increasing alarm. 

In Zhang v. Chen (2024 BCSC 285), a Vancouver lawyer filed a custody application citing two cases that didn't exist. She'd pulled them from ChatGPT without verification. When opposing counsel couldn't locate them and confirmed they were fabricated, Justice Masuhara ruled that "generative AI is still no substitute for the professional expertise that the justice system requires of lawyers." The Law Society of BC launched an investigation. 

In Ontario's R. v. Chand (2025 ONCJ 282), Justice Kenkel found "numerous and substantial" errors in defence counsel's submissions for an aggravated assault case, including at least one fictitious case and several citations to unrelated civil cases in a criminal proceeding about self-defence. The judge ordered entirely new submissions and explicitly prohibited the use of "generative AI or commercial legal software that uses GenAI" for legal research going forward. 

These aren't embarrassed lawyers afraid to admit they're using helpful tools. These are lawyers who delegated professional judgment to a system incapable of exercising it. 

The LSO's April 2024 White Paper on AI states this directly: "Licensees cannot inappropriately delegate tasks. Certain tasks require the input of a licensee and the application of their legal competence. Generative AI is not a replacement for this." 

Here's what Raskin doesn't understand about what lawyers actually do. Robert Diab from Thompson Rivers University puts it plainly: "AI won't be able to hold the client's hand, negotiate with the other side, use emotional intelligence to read the room, or to read the client and to formulate the right questions." 

Legal work isn't about knowing the rule against perpetuities. It's about recognizing when your client is hiding something that will destroy their case. It's about reading a room during a mediation and knowing when to push and when to back off. It's about understanding that the technically correct legal strategy will devastate your client emotionally and finding a better path. 

That's judgment. And Kim Kardashian's own experience demonstrates why it matters that we assess for it. 

Kardashian is the argument for meaningful licensing 

The irony of Raskin's editorial, and he seems entirely unaware of it, is that Kim Kardashian validates rather than undermines the case for licensing standards. 

Kardashian revealed in her announcement that she'd been using ChatGPT to help study for the exam. "It made me fail tests," she said, "all the time" because it gave her wrong answers. 

Think about what that means. A person preparing to practise law, to advise vulnerable people on matters affecting their liberty, their families, their livelihoods, couldn't recognize when an AI system was giving her incorrect legal information. She lacked the foundational judgment to evaluate the tool she was relying on. 

That's not an indictment of her intelligence. It's evidence that legal competence involves something more than can be acquired through celebrity apprenticeship and AI study aids. 

Raskin dismisses this concern with a wave. The market will sort it out. Bad lawyers will get bad reviews. People will learn to hire better representation. 

But legal errors don't work like bad Yelp reviews for a restaurant. By the time you discover your lawyer was incompetent, you may have lost custody of your children, been wrongfully convicted, or signed away rights you didn't understand you had. The harm is irreversible. The information asymmetry between lawyers and clients means people can't evaluate quality in advance. 

This is why every developed legal system, not just common law jurisdictions, requires licensing. It's not a cartel. It's a recognition that legal services affect fundamental rights in ways that market mechanisms cannot efficiently correct after the fact. 

Ontario shows what regulated innovation really looks like 

Raskin's access-to-justice argument deserves to be taken seriously. Legal costs have risen at roughly twice the rate of inflation over the past decade. The 2016 Bonkalo Report found that 57% of family court litigants in Ontario lacked legal representation. 

But here's where his analysis becomes intellectually lazy: he assumes the only response to access barriers is deregulation. Ontario's experience shows otherwise. 

In May 2007, Ontario became the first North American jurisdiction to license paralegals. These aren't lawyers-lite operating in a grey market. They're professionals who complete accredited education, pass licensing examinations, maintain professional liability insurance, and submit to the same disciplinary oversight as lawyers. 

Paralegals can now handle small claims, landlord–tenant disputes, traffic violations, and administrative tribunals. This expanded access to legal services while maintaining accountability. 

In December 2022, the LSO approved the Family Legal Services Provider program, allowing specially trained paralegals to provide limited family law services after completing a 260-hour program through Fanshawe College. This directly addresses the family court access crisis within a regulatory framework. The first cohort started in January 2025. 

In November 2021, the LSO launched its Access to Innovation sandbox, a five-year pilot allowing providers of innovative technological legal services to operate under supervision. As of May 2024, 11 approved participants have provided over 50,000 legal services with only four complaints. This creates space for legal technology innovation while maintaining public protection conditions and accountability mechanisms. 

These initiatives demonstrate that the choice isn't binary. Ontario has expanded who can provide legal services, what technology can be deployed, and how services can be structured, all while maintaining the regulatory accountability that protects vulnerable clients. 

What you're actually buying when you hire a regulated lawyer 

Raskin complains about restrictions on non-lawyers as law firm partners, arguing this prevents beneficial specialization. "Why would someone think that a lawyer who has trained in a narrow field would be good at firm operations or marketing or hiring?" he asks. 

It's a reasonable question with a straightforward answer: law firms provide legal services, and legal services require professional judgment that must remain independent of commercial pressures. 

When you hire a lawyer in Ontario, you're not just buying legal knowledge. You're buying: 

• Privilege: Communications with your lawyer are protected. Communications with an AI chatbot or unregulated service provider are not. 

• Insurance: All Ontario lawyers in private practice must maintain $1 million per claim coverage through LAWPRO. If your lawyer makes a catastrophic error, you have recourse. 

• Disciplinary oversight: The Law Society Tribunal can investigate complaints and impose sanctions ranging from fines to disbarment. The LSO addressed 91% of complaints within 12 months in 2024, exceeding national standards. 

• Ongoing competence requirements: Ontario lawyers must complete 12 hours of continuing professional development annually, including three hours on professionalism with mandatory equality, diversity, and inclusion content. The LSO conducts practice audits, with mandatory audits for all newly formed practices within their first 12 months of operation. 

• Ethical obligations: Lawyers must avoid conflicts of interest, maintain client confidentiality, and exercise independent professional judgment even when it conflicts with their own financial interests. 

This infrastructure exists not at the point of initial licensing but throughout a lawyer's entire career. The bar examination is one component of a comprehensive regulatory system, not its totality. 

The real Kim Kardashian story 

Let's return to where we started. Kim Kardashian failed the bar exam. So did Kamala Harris on her first attempt. So did countless capable people who went on to distinguished legal careers. 

The exam is a flawed measure. It needs reform. That's the story Raskin should have written. 

Instead, he used Kardashian's failure to argue for dismantling professional regulation entirely. The problem is that Kardashian—a celebrity with unlimited resources, years to study, and access to any tutor she wants—demonstrates the opposite of what Raskin claims. 

If someone with her advantages struggles to master legal fundamentals and distinguish good information from AI hallucinations, what does that tell us about the complexity of legal competence? What does it suggest about the risks of unregulated practice? 

The answer shouldn't be to make it easier for unprepared people to hold themselves out as lawyers. It should be to make licensing better at identifying and developing the judgment, ethics, and skills that clients need. 

Why this matters beyond one celebrity's exam results 

There's a broader point here about how we think about professional regulation in an age of technological disruption. 

Raskin's argument rests on a libertarian assumption that markets are efficient information processors and that regulation primarily serves to protect incumbents. Sometimes that's true. Often, it's not. 

Professional regulation in law serves a public protection mandate that markets cannot replicate. The information asymmetry between lawyers and clients, the irreversibility of many legal harms, and the fundamental rights at stake all justify regulatory oversight that goes beyond what market signals alone provide. 

But meaningful regulation has to justify itself through competence, not just tradition. It has to reform flawed practices, embrace beneficial innovation, and demonstrate that the costs of oversight produce genuine public benefit. 

We can acknowledge that traditional bar exams are imperfect while maintaining that professional judgment, ethical accountability, and regulatory oversight serve the public in ways that neither markets nor machines can replicate. 

The real failure here isn't Kim Kardashian's bar exam result. It's Max Raskin's failure to understand the difference between a test that needs reform and a profession that needs dismantling. 

Ontario knows better. And lawyers who care about access to justice while protecting vulnerable clients should pay attention to what we're building here: better regulation, not deregulation.