We only use ChatGPT. Does the AI Act really apply to us?+
Yes. Using an AI system at work makes you a deployer. An AI-literacy duty has applied since 2 February 2025 (its current wording comes from the AI Omnibus), and the ban on prohibited practices applies too. For a general assistant the rest is light, and the checker shows exactly what.
What do we have to do right now?+
Two duties already apply: take measures for AI literacy among staff who use AI (Art. 4), and make sure you don't use any banned practice (Art. 5). Since 2 August 2026, chatbots and AI-generated content also need transparency notices (Art. 50). The free check shows which of these apply to your tools and what is due next.
When do the high-risk rules start?+
After the AI Omnibus, on 2 December 2027 for stand-alone uses such as hiring and credit scoring, and on 2 August 2028 for AI built into regulated products such as medical devices or lifts.
Do AI agents we built ourselves fall under the AI Act?+
Yes. If you build an agent and use it in your company, you are usually its deployer and can also be its provider (Art. 3(3)–(4)). Most internal agents carry light duties: AI literacy (Art. 4), no banned practices (Art. 5) and transparency where people interact with the agent or receive AI-generated content (Art. 50). An agent used for decisions about staff, such as allocating tasks or evaluating performance, is high-risk (Annex III point 4). Our Internal AI Agent Review checks up to three agents.
What do you need from us for a bias test?+
Past outcomes of the system, such as which applicants it shortlisted or how it scored them; the group characteristics to compare, such as gender or age band; and a legal basis for using that data under the GDPR. We sign a data processing agreement and work with pseudonymised data where we can. The test takes about a week once the data arrives. More on the bias test.
Do we need to involve our works council?+
In Germany, often yes, if you have one. The works council has a right of co-determination when a technical system is suited to monitor employees' behaviour or performance (§ 87(1) no. 6 BetrVG), which many AI tools are. It must be informed and consulted when you plan to use AI in work processes (§ 90 BetrVG), has a say in selection guidelines also when AI is used (§ 95(2a) BetrVG) and may call in an expert on AI (§ 80(3) BetrVG). Plan for this early.
We have SOC 2. Does that cover the AI Act?+
No. SOC 2 is a voluntary attestation, under AICPA criteria, about controls such as security and confidentiality. The AI Act is EU law, and its duties depend on your role and the risk class of each AI system: AI literacy, banned practices and transparency for everyone, and for high-risk systems risk management, data governance, documentation, logging and human oversight. A SOC 2 report helps with parts such as cybersecurity, but it does not show AI Act compliance.
Do you give legal advice?+
No. GenAIze Technologies UG (haftungsbeschränkt) provides technical and organisational consulting on the use of AI, and information and self-assessment software that applies general rules to your answers and cites the article behind every legal conclusion. Neither is legal advice. When you need individual legal advice, please consult a qualified lawyer.
Who is behind GenAIze?+
GenAIze Technologies UG (haftungsbeschränkt) in Berlin, led by Managing Director Aqsa Peer, who holds the A4Q AI Compliance certificate, a private personnel certification of A4Q issued by GASQ Service GmbH (certificate no. 8548357428, valid until 5 August 2028). She is not a lawyer; for individual legal advice, please consult a qualified lawyer.
Where is our data stored?+
The free checker runs in your browser and sends nothing to us. Share links keep your answers inside the link itself, and downloads are created on your device. If you join the early-access list, we receive your email address and, only if you choose, a short summary of your check. If you sign in to the Academy, we store your email address and the modules you passed, to issue your certificates. Our email service Brevo processes this data for us, and our website host Netlify runs the site and marks quiz answers without storing them, both under data processing agreements. In our services, we use your data only for the engagement and delete it after delivery; an NDA and a data processing agreement are available. Details are in our privacy policy.
How is this different from a GRC platform?+
GRC tools assume a compliance team that feeds them. GenAIze starts from the AI tools a small company actually uses: a free check in plain language, with the article behind every result; free training for your staff; and draft documents you can use straight away. It is built for companies without a compliance department.
Is it really free?+
Yes. The checker, the Academy, the certificates, the downloads and the draft documents are free, with no sign-up for the checker. Our services are paid, with prices on request, and the first call is free. You don't need a service to use the free tools.