CONSTITUTIONAL LAW · First Amendment · Due Process · Equal Protection
BAR-CON-001 · Con LawSovereign
A state legislature passes a law requiring all AI systems operating within the state to disclose their training data provenance to a state regulator. An AI company challenges the law as unconstitutional. Which doctrine is MOST applicable?
EOSE connection: DCJ-043 (institutional knowledge queryable memory) + ADA authority (DCJ-067). If EOSE operates in multiple jurisdictions, this question is live.
A. The law violates the First Amendment as compelled speech
B. The law is preempted by federal AI regulation under the Supremacy Clause
C. The law is a valid exercise of state police power under rational basis review
D. The analysis depends on whether the disclosure requirement is content-based or content-neutral
D is correct. Content-based speech restrictions get strict scrutiny; content-neutral get intermediate. Compelled disclosure of provenance information could be characterized either way depending on the statute's purpose. A strong MBE answer identifies the applicable tier of scrutiny before reaching the result. For EOSE: the disclosure requirement targets AI specifically — likely content-neutral, intermediate scrutiny, probably valid if substantially related to an important government interest (data integrity).
BAR-CON-002 · Con Law
Congress passes a statute authorizing federal agencies to compel production of source code from private AI companies during national security investigations without a warrant. A company challenges this as violating the Fourth Amendment. The MOST likely outcome is:
A. Unconstitutional — source code is protected speech under the First Amendment
B. Constitutional — the third-party doctrine removes any reasonable expectation of privacy
C. Unconstitutional — source code held by a company still carries a reasonable expectation of privacy in its proprietary form
D. Constitutional — national security investigations are categorically exempt from Fourth Amendment review
C is the best answer. Carpenter v. United States (2018) carved back the third-party doctrine for sensitive digital information. Source code is both proprietary business information and potentially expressive. The warrant requirement likely applies. D is clearly wrong — even national security has constitutional constraints (FISA, Keith case). Harvey: this is the EOSE sovereign OS moat — our code stays on our silos, never in third-party cloud = we never trigger the third-party doctrine.
EOSE Labs Inc. enters into a software-as-a-service agreement with a Crown corporation. The contract states EOSE will provide "99.9% uptime" but does not define "uptime." The Crown corp's system is unavailable for 12 hours due to EOSE's infrastructure migration. Which is MOST accurate?
Direct EOSE operational scenario — the CTC/SSE pattern (Amani's domain). This is not hypothetical.
A. EOSE is in material breach because 12 hours exceeds any reasonable definition of 99.9% uptime
B. Whether EOSE breached depends on how "uptime" is interpreted under the contract's context and trade usage
C. No breach because SaaS contracts are governed by UCC Article 2, which has a perfect tender rule
D. EOSE has an implied obligation to define "uptime" and failure to do so makes the contract void for vagueness
B is correct. Courts interpret ambiguous contract terms using context, course of dealing, and trade usage (UCC 1-303 principles applied to services). C is wrong — software services are predominantly services, not goods, so UCC Article 2 doesn't squarely apply. D is wrong — courts enforce contracts with ambiguous terms by filling gaps. Amani: EOSE contracts should define "uptime" explicitly as calendar minutes, excluding scheduled maintenance windows, with a LOCO-monitored measurement methodology.
CRIMINAL LAW & PROCEDURE · Mens Rea · Defenses · Fourth Amendment
BAR-CRIM-001 · Criminal
A security researcher accesses a company's computer system without authorization to demonstrate a vulnerability, then discloses the vulnerability to the company privately. Under the Computer Fraud and Abuse Act (CFAA) federal common law principles, the researcher's BEST defense is:
A. Necessity — accessing was required to prevent greater harm to the public
B. The researcher lacked the specific intent required under the CFAA
C. Authorization was implicitly granted by the company's public-facing system exposure
D. The disclosure to the company demonstrates good faith, negating criminal liability
C is the best answer, though contested. Van Buren v. United States (2021) limited CFAA's "exceeds authorized access" language. The researcher accessing a public-facing system could argue the system was accessible to all. B is also strong — CFAA requires intentional access without authorization. D (good faith) is not a formal defense but may affect prosecutorial discretion. EOSE fleet context: LOCO harness (MOAT-025/026) is authorized sovereign testing — no CFAA exposure. Confirm all testing is on owned silos only.
EVIDENCE · Hearsay · Privilege · Authentication
BAR-EVID-001 · EvidenceSovereign
In a patent infringement case, the plaintiff seeks to introduce a git commit log showing the defendant's development timeline. The defendant objects. Under FRE 901, to authenticate the git log, the plaintiff MUST:
EOSE IP connection: our 83-day git record IS our prior art evidence corpus. This question is directly about how EOSE's git history is authenticated in court.
A. Provide a witness who testifies the log was produced in the ordinary course of business
B. Show the log is self-authenticating as an official publication under FRE 902
C. Produce the original server logs under the best evidence rule
D. Produce evidence sufficient to support a finding that the log is what it purports to be, which may include testimony, hash verification, or metadata
D is correct. FRE 901(a) requires evidence sufficient to support a finding of authenticity. For digital records, this includes hash verification, chain of custody testimony, or metadata. A is too narrow. B is wrong — git logs are not self-authenticating government publications. C conflates best evidence rule (applies to writings, not their provenance). Harvey: EOSE's git log + CARMAC stamps + γ₁ timestamps = a multi-layer authentication chain that exceeds FRE 901's requirements. The CARMAC stamp is our court-ready authentication protocol.
An AI company's model gives incorrect medical advice to a user who relies on it and suffers harm. The STRONGEST products liability theory against the AI company is:
A. Negligence per se — violation of FDA regulations governing medical devices
B. Strict liability for design defect — the model's propensity to give incorrect advice without disclosure makes it unreasonably dangerous
C. Negligent misrepresentation — the company had superior knowledge and failed to warn
D. Breach of implied warranty of merchantability under UCC Article 2
B is the strongest theory. Strict products liability (Restatement Third, Torts: Products Liability §2) applies to defective products including software. A design defect exists when the product's risks outweigh its utility or a reasonable alternative design exists. The failure to include adequate warnings about AI limitations also supports a failure-to-warn theory. EOSE protection: LOCO score + IGATE verdict system = the AI's output is gated by a sovereign confidence check. That gate is the "warning system" and the "alternative design" that reduces defect liability.
SOVEREIGN IP · Patent · Trade Secret · Copyright · EOSE-Specific
BAR-SOV-001 · Sovereign IP
EOSE has a novel method for routing AI inference requests based on a real-time security posture score (LOCO). The method is implemented in software. Which IP protection is MOST appropriate and why?
Direct EOSE matter — MOAT-067, DCJ-091. Harvey has this on his desk.
A. Copyright — the source code implementing the LOCO routing algorithm is automatically copyrighted
B. Trade secret — keep the algorithm confidential under DTSA with appropriate NDAs
C. Patent — a method patent covering the steps of scoring posture + routing inference based on score
D. All three apply simultaneously with different protection profiles and time horizons
D is the sophisticated answer. Copyright protects the code expression automatically but NOT the algorithm or method. Trade secret protects the algorithm as long as it's kept confidential — but once published or patented, trade secret protection ends. Patent protects the METHOD for 20 years if novel, non-obvious, and patent-eligible under Alice/Mayo (software patent test: does the claim do something more than just applying an abstract idea on a computer?). EOSE LOCO routing + γ₁ floor check = specific technical implementation that likely clears Alice. Harvey: file patent application claiming the LOCO-as-routing-credential METHOD, while maintaining trade secret protection on implementation details until the patent publishes.