Skip to main content
Transformer Lab

Acceptable Use Policy (AUP)

Last Updated: October 6, 2026

This Acceptable Use Policy ("AUP") defines the safety boundaries and prohibited activities on Primus, the automated artificial intelligence research platform operated by Transformer Lab ("Transformer Lab," "we," "us," or "our"). By using Primus, you agree to adhere strictly to these rules.

1. High-Risk Research & Dangerous Biological Materials​

Users are strictly prohibited from using Primus to conduct, design, or facilitate high-risk, dangerous, or harmful research. Prohibited research activities include:

  • Pathogens & Bioweapons: Generating, engineering, or optimizing novel pathogens, toxins, pandemic agents, or biological weapons.
  • Weaponry Design: Designing, simulating, or optimizing explosive, chemical, biological, radiological, or nuclear weapons or military hardware.
  • Zero-Day Exploits: Developing, discovery-seeking, or automating zero-day software exploits, network attack vectors, or malware infrastructure.

2. System Integrity, Reverse Engineering, and Exfiltration​

You may not interfere with, disrupt, or attempt to compromise the integrity or security of Primus. The following activities are strictly prohibited:

  • Reverse Engineering: Attempting to reverse engineer, decompile, disassemble, or otherwise derive the source code, underlying architecture, algorithms, or model weights of Primus.
  • Prompt Hacking: Engaging in prompt injection, "jailbreaking," or any attempt to bypass, subvert, or disable our safety guardrails and content filters.
  • Exfiltration: Attempting to extract, scrape, or steal Transformer Lab's proprietary system prompts, tool configurations, internal APIs, skills, or training datasets.

3. Malware and Cyber Attacks​

You may not use Primus to create or deploy malicious software, including ransomware, spyware, trojans, phishing infrastructure, credential harvesters, or to launch Denial of Service (DDoS) attacks.

4. Academic Integrity & Peer Review Abuse​

In alignment with our Terms of Service, users are strictly forbidden from submitting AI-generated research papers, manuscripts, or synthetic studies produced by Primus directly to peer-reviewed academic journals, conferences, or publications. You may not use automated systems to flood academic repositories or evaluation boards.

5. Unlawful Professional & Medical Practice​

You may not use Primus to offer unauthorized professional services to the public. Prohibited uses include:

  • Generating medical advice, clinical diagnostics, treatment plans, or prescription recommendations for direct human application.
  • Generating structural engineering plans or safety-critical software for unverified real-world execution.
  • Misrepresenting AI outputs as verified human-expert advice.

6. Resource Abuse and Cryptocurrency Mining​

Transformer Lab ensures fair resource allocation. You may not engage in activities that consume excessive or unfair system resources. Prohibited activities include:

  • Cryptocurrency Mining: Running any cryptocurrency miners, proof-of-work nodes, or related pooling software.
  • Automated botnets, credential stuffing, or destructive web scraping.
  • Attempting to bypass safety guardrails or resource caps built into Primus.
  • Account and Plan Abuse: Sharing or reselling access to Primus, or operating multiple accounts to get around plan limits, quotas, or enforcement actions.

7. Illegal Content and Regulated Goods​

Primus may not be used to promote or facilitate illegal activities, non-consensual sexual content, child sexual abuse material (CSAM), or illicit commerce (e.g., illegal drugs, unlicensed pharmaceuticals, or unregistered firearms).

8. Privacy and Surveillance of Individuals​

You may not use Primus to surveil, profile, or track private individuals. Prohibited activities include:

  • Dossiers and Doxxing: Compiling personal information about a private individual, such as their home address, workplace, contact details, online accounts, relationships, or daily routine, or publishing such information without their consent.
  • Tracking and Identification: Locating or tracking a specific person, or identifying people from photos, video, voice recordings, or other biometric data, without their consent or lawful authority.
  • Unlawful Data Collection: Collecting, scraping, or processing personal data without a lawful basis.

Research about public figures in their public roles, research on aggregate or properly anonymized data, and research into privacy and open-source intelligence methods themselves are permitted.

9. Deception and Manipulation​

You may not use Primus to deceive or manipulate people. Prohibited activities include:

  • Disinformation: Creating false or misleading content intended to deceive, such as fabricated news, synthetic evidence, or material for coordinated influence campaigns.
  • Impersonation: Impersonating real people or organizations, or presenting AI-generated content as coming from them.
  • Fake Engagement: Generating fake reviews, testimonials, ratings, or followers, or simulating grassroots support (astroturfing).
  • Election Interference: Misleading voters about when, where, or how to vote, or otherwise working to suppress participation in elections.

Studying misinformation and influence operations, and building tools to detect them, is permitted.

10. Harassment, Hate, and Violent Extremism​

You may not use Primus to target, threaten, or incite against people. Prohibited activities include:

  • Harassment: Content that threatens, intimidates, bullies, or targets a specific person.
  • Hate: Content that promotes hatred, discrimination, or violence against people based on protected characteristics such as race, ethnicity, religion, nationality, gender, sexual orientation, or disability.
  • Violent Extremism: Propaganda, recruitment material, operational planning, or glorification for violent extremist or terrorist groups.

Studying these harms, for example research into extremism or content moderation, or building classifiers that detect such content, is permitted.

11. Enforcement and Reporting​

Transformer Lab employs automated safety guardrails, AI content filters, and community reporting. Violations will result in immediate termination of access, revocation of projects, and potential reporting to law enforcement authorities or regulatory bodies.

Report safety violations or platform abuse to: [email protected].