AI Use Policy

GBCORP develops and operates artificial-intelligence products and services as one of our six business pillars. This AI Use Policy describes acceptable and unacceptable use of GBCORP AI capabilities, our approach to data, and our limitations.

1) Scope

This Policy covers all AI-powered products and services operated by GBCORP, including:

• GBCORP AI APIs and SDKs (gbcorp.ai)
• AI features inside GBCORP business services (Voice AI, hyperautomation, generative AI tools)
• The GBCORP AI Humanoid Robot platform and software
• AI integrations within IT, business-services, and education products

2) Permitted Uses

You may use GBCORP AI services for lawful, commercial, educational, or research purposes consistent with your subscription, the underlying agreement, and this Policy. Typical permitted uses include:

• Productivity automation, document processing, and workflow orchestration
• Customer-service assistants and internal knowledge-base tools
• Content generation for review and editing by humans before publication
• Research, prototyping, and product development
• Educational use within a learning environment

3) Prohibited Uses

You may NOT use GBCORP AI services to:

• Harm people — generate weapons-of-mass-destruction information, create violent or graphic content, plan attacks, or facilitate self-harm
• Deceive — create deepfakes, impersonations, fraudulent documents, manipulated media, or disinformation campaigns
• Violate privacy — perform facial recognition or biometric inference on non-consenting individuals; surveil populations; doxx individuals
• Discriminate unlawfully — make automated decisions affecting credit, employment, housing, insurance, or essential public services without lawful basis, human review, and appropriate safeguards
• Generate CSAM — child sexual abuse material is absolutely prohibited and will be reported to authorities
• Generate political manipulation content — fake endorsements, fake political ads, vote-suppression messaging, or coordinated inauthentic behaviour
• Infringe intellectual property — train derivative models without licence, generate content that copies copyrighted works, or pass off AI-generated work as covered by a third-party rights-holder's licence
• Bypass safety — jailbreak, prompt-inject, or otherwise attempt to circumvent the safety classifiers and content filters in GBCORP AI services
• Reverse engineer — extract model weights, training data, or other proprietary information through interrogation, model inversion, or distillation

4) Data Handling and Training

By default:

• We do not train our foundation models on Customer prompts, outputs, or uploaded content without explicit written consent
• Input and output data may be retained for short periods to operate the Service (default 30 days) and may be reviewed by GBCORP staff only as necessary to investigate abuse or to comply with law
• Enterprise customers can configure shorter retention or zero-retention on request
• Public training data used to build our foundation models is sourced from publicly available datasets and licensed content; our data-source registry is available to enterprise customers on request

5) Output Accuracy and Reliability

AI models can produce inaccurate, incomplete, biased, or fabricated content ("hallucinations"). You are responsible for:

• Reviewing all AI-generated output before relying on it in a real-world decision
• Not using AI-generated output as the sole basis for high-stakes decisions (medical, legal, financial, employment, safety-critical)
• Clearly disclosing to end users when content is AI-generated, where required by law (e.g., EU AI Act, EU Digital Services Act, state and national synthetic-content laws)

6) Human Oversight

For high-risk applications (as defined under the EU AI Act and equivalent frameworks), Customers must maintain meaningful human oversight, including the ability to intervene, override, and audit AI decisions.

6.1) Regulatory Alignment — EU AI Act

GBCORP AI offerings are categorised and disclosed in line with the EU AI Act (Regulation (EU) 2024/1689), including the Article 50 transparency obligations applicable from 2 August 2026 (synthetic content disclosure, AI-system interaction notices, and emotion-recognition / biometric-categorisation notices where relevant). Customers deploying GBCORP AI in scope of the Act remain responsible for fulfilling their own provider, deployer, or distributor obligations.

7) Bias and Fairness

GBCORP tests its AI models for known biases and continues to improve them. Where bias is found that materially affects Customer use, we will notify affected Customers and provide mitigations where reasonably possible.

8) Intellectual Property

Subject to your underlying agreement and any input-data restrictions:

• You retain rights in your input prompts and Customer-provided data
• You receive a licence to use the output of the Service for any lawful purpose, subject to this Policy
• GBCORP retains all rights in its models, infrastructure, and proprietary technology

9) Reporting Concerns

Report AI safety concerns, harmful output, or suspected policy violations to:

• [email protected] for safety and abuse
• [email protected] for jailbreaking, prompt injection, or model-extraction reports (covered by our Vulnerability Disclosure Policy)

10) Updates

This Policy will evolve as the regulatory landscape and GBCORP's AI capabilities mature. Material changes will be notified to active Customers with at least 30 days' notice.