EU AI Act — Regulatory Compliance

Under the EU AI Act, AskBiz operates as a deployer — an entity that uses an AI system (Claude by Anthropic) in the course of a professional activity to provide a service to users.

Anthropic operates as the provider of the Claude general-purpose AI model. Our compliance obligations as deployer are set out in Article 26 of the AI Act.

As a deployer of a limited-risk AI system, AskBiz fulfils the following Article 26 obligations:

Use in accordance with instructions: We use Claude within the scope of its intended purpose as defined by Anthropic, in accordance with Anthropic's usage policies.

Human oversight: We implement appropriate human oversight measures. Our interface presents AI outputs with confidence indicators and encourages human review. The AI cannot take autonomous consequential actions.

Data governance: We ensure the input data we use to query the AI is relevant, sufficiently accurate, and appropriate for the intended purpose.

Monitoring: We monitor the AI system's operation for risks and report concerns to Anthropic through established channels.

Transparency to users: We inform users that they are interacting with an AI system and provide the AI Transparency disclosures required by Article 50.

AskBiz maintains the following documentation in compliance with the AI Act:

• Risk classification assessment — documenting why AskBiz is classified as limited-risk
• Deployer technical documentation — describing the AI system, its inputs/outputs, and our use case
• Sub-processor agreements — including data processing terms with Anthropic
• Incident log — record of any AI-related incidents or near-misses
• User rights log — record of user requests to explain or contest AI decisions
• Annual review record — annual assessment of continued compliance

This documentation is available to EU supervisory authorities on request.

The AI Act's Article 5 prohibitions on unacceptable-risk AI practices apply from 2 February 2025. AskBiz complies:

• We do not use AI for social scoring of individuals
• We do not use AI for subliminal manipulation
• We do not use real-time biometric identification systems
• We do not use AI for predictive policing or pre-crime assessment
• We do not use AI to exploit vulnerabilities of specific groups
• We do not use AI to infer political, religious, philosophical beliefs, or trade union membership
• We do not use AI emotion recognition in workplace or educational settings

These commitments apply regardless of user requests.

As an EU user, you have rights related to AI under the AI Act:

Right to explanation: Ask any AskBiz AI response to explain its reasoning. The AI is instructed to answer these meta-questions transparently.

Right to contest: If an AI output influences a decision of significance to you and you believe it is incorrect, email support@askbiz.co with the context. A human will review.

Right to complain: If you believe AskBiz has violated the AI Act, you may complain to the national supervisory authority in your EU member state. In Ireland, this is the AI Office (established under the AI Act). In other member states, the relevant authority is typically the data protection authority or a designated AI authority.

Contact: support@askbiz.co for all AI-related queries.