Laws & Policy
EU AI Act
The European Union law that sorts AI by risk, with stricter duties for some high-risk uses, including listed education systems.
As of September 26, 2026, transparency rules apply. Annex III high-risk duties, including listed education uses, apply from December 2, 2027, after Regulation (EU) 2026/1744. Check the European Commission page before you rely on these dates.
- Policymakers and staff
- Businesses
This is general information, not legal advice.
What parents should know
The EU AI Act sorts AI by risk. Some uses are banned. Some must be transparent, including chatbots. Some education tools are labeled high-risk, and those fuller duties were delayed to December 2, 2027. Dates move. Read the Commission page, not a 2024 slide. This is general information, not legal advice.
On this page
What is the EU AI Act?
Regulation (EU) 2024/1689 entered into force on August 1, 2024. It does not treat every chatbot the same. Prohibited practices and AI literacy duties have applied since February 2, 2025. Duties for general-purpose models applied from August 2, 2025. Transparency duties, including telling people they are talking to a chatbot, applied from August 2, 2026.
High-risk uses listed in Annex III include some education systems, such as tools that decide admission, evaluate learning outcomes, or monitor students during tests. Those Annex III duties were originally due August 2, 2026. Regulation (EU) 2026/1744, in force July 27, 2026, moved them to December 2, 2027. High-risk AI embedded in regulated products was moved to August 2, 2028. A timeline that still says August 2026 for Annex III is out of date.
Why the EU AI Act matters
A US family is not the Act's main reader, and a US company that offers AI to people in the EU can still be in scope. Schools shopping for a European vendor will see the education category and should ask which date the vendor is actually building toward.
The child-specific piece is easy to overstate. The Act restricts practices that exploit vulnerabilities related to age, and it treats some education uses as high-risk. It is not a single kids' mode switch. Ask which article the claim is about, and whether that article is already in force.
How it shows up in practice
- A business checks whether a chatbot disclosure duty already applies, separate from the later high-risk paperwork.
- A school leader asks a vendor if an exam or admissions tool is an Annex III system and which deadline it is using.
- A policy staffer throws out a briefing that still lists August 2, 2026 as the high-risk start.
- A parent in the EU asks whether a companion-style bot has to say it is a machine.
How HeyOtto helps
HeyOtto does not claim to be certified under the EU AI Act. Otto is presented as the assistant, not a person. A parent creates the family account, can read the chats, and can set topic limits. Businesses that want to offer a parent-visible kids' experience can talk with HeyOtto about a partnership. That conversation is not an EU conformity assessment.
- Chat is visibly the assistant. It is not introduced as a human friend.
- Schools get teacher and advisor dashboards and can turn student tools on or off.
- Do not treat this page as a legal opinion on whether a deployment is in scope.
For businesses
Talk with HeyOttoFAQs
Is the EU AI Act fully in force?
No. It applies in stages. As of September 26, 2026, prohibited practices, AI literacy, general-purpose model duties, and transparency duties are in application. Annex III high-risk duties, including listed education uses, apply from December 2, 2027. Product-embedded high-risk systems apply from August 2, 2028. Confirm dates on the Commission page.
Are kids' chatbots high-risk?
Not automatically. Annex III names specific education uses, such as admissions, evaluating learning outcomes, assigning a level of education, and monitoring tests. A general chat is more often a transparency question: people should know they are talking to a machine. Read the annex before you apply the high-risk label.
Did the 2026 delay cancel the Act?
No. Regulation (EU) 2026/1744 deferred the high-risk deadlines. It did not repeal the Act. Transparency rules were not pushed to 2027. If a slide still shows August 2, 2026 for Annex III, it predates the omnibus.
Does this replace COPPA for a US app?
No. COPPA is US law for personal information from children under 13. The EU AI Act is a European product-safety and fundamental-rights law. A service can face both if it operates in both places. This page is general information, not legal advice.
Sources
Last reviewed September 26, 2026. Law entries are reviewed every quarter.
This is general information, not legal advice.