Skip to content
Lacambra Web Portal Services
← 2026, week by week
Regulation · Week 31, 27 July to 2 August 2026

EU AI Act from 2 August: what reaches Philippine firms

The EU AI Act applied from 2 August 2026, parts of it delayed days before. It reaches firms whose AI output is used in Europe, Philippine ones too.

Three plain cream envelopes in a row on a dark wooden desk, each closed with a wax seal, the first seal broken open and the other two still whole.

On 2 August 2026 the European Union's AI Act became applicable, with some exceptions. Six days earlier an amending regulation, the AI Omnibus, had entered into force and moved the rules for high-risk uses, such as screening job applicants or scoring credit, to 2 December 2027. The transparency rules did not move: from 2 August, people must be told when they are dealing with an AI, and deepfakes must be labelled. Those rules also reach Philippine firms whose AI work is used in Europe.

What applied on 2 August, and what moved

The Act entered into force on 1 August 2024 and has arrived in stages, the European Commission sets out: its bans and AI literacy duty from February 2025, the rules for general-purpose models from August 2025, the rest from 2 August 2026. The Omnibus, Regulation (EU) 2026/1744, agreed in May and published in the Official Journal on 24 July, pushed back the high-risk rules in Annex III, which covers, among other uses, AI used to recruit people, evaluate staff or judge creditworthiness, to 2 December 2027, and those for AI built into products such as lifts and toys to 2 August 2028.

Article 50, on transparency, kept its date. A system built to talk with people must tell them it is an AI, unless that is obvious, by the first interaction at the latest. Systems that generate images, audio, video or text must mark their output in a machine-readable way, and a business that uses AI to make a deepfake must disclose it. Generative systems already on the market before 2 August have until 2 December 2026 for the marking.

Why it reaches a Philippine firm

The reach is in Article 2. The Act covers providers that put AI systems on the EU market wherever they are established, and providers and deployers in a third country "where the output produced by the AI system is used in the Union". A Philippine contact centre whose AI chatbot answers a European client's customers, or a software house that builds an AI feature for a client in the EU, produces output used in the Union, and can be inside the Act.

The fines are set in Article 99: breaking the transparency rules can cost up to €15 million or 3% of worldwide annual turnover, whichever is higher, and for small and medium-sized firms whichever is lower.

What it means for a supplier with clients abroad

Serving European clients from here is ordinary: I have clients in eight countries, among them a spare-parts retailer in Germany. If a system built here for a European client includes AI, the Act's duties come with it.

Who carries them is the next question. Article 50 splits its duties between the provider, which Article 3 defines as whoever develops an AI system, or has one developed, and puts it into service under its own name, and the deployer, whoever uses it under its authority. On a system built to order and run under the client's name, which party holds which role is for the contract, and a lawyer, to settle before the build. Either way, the notice and the marking are built by whoever writes the code.

The delay on high-risk uses is time to design, not time off. Any AI that screens applicants, evaluates staff or scores credit should work the way a good approval chain does: one named person owns each decision, and the record shows what the system suggested and what the person decided.

Five checks before 2 December

  • List every place AI touches work whose output reaches people in the EU: chat on a client's site, drafted replies, generated images, screening for roles in Europe.
  • Anything that talks to people should say it is an AI at the start. Build the notice into the system, not into a policy.
  • Check that any tool generating images, audio, video or text marks its output, and label deepfakes. Systems on the market before 2 August have until 2 December 2026.
  • Write into each contract who is the provider, who is the deployer and who builds the notices. Have a lawyer read that clause.
  • Support your staff's AI literacy, as Article 4 asks, with a short guide to the tools in use and where they go wrong. Since the Omnibus, no set level is required.

If you are about to put AI in front of people in Europe and want the notices built in from the first version, the fit check is the place to start.

Sources

  1. European Commission, AI Act, 19 October 2026
  2. The National Law Review, EU Digital Omnibus on AI Enters Into Force, 31 July 2026
  3. Gibson Dunn, EU AI Act Omnibus Agreement: Postponed High-Risk Deadlines and Other Key Changes, 27 May 2026
  4. AI Act Service Desk, European Commission, Article 2: Scope
  5. AI Act Service Desk, European Commission, Article 3: Definitions
  6. AI Act Service Desk, European Commission, Article 4: AI literacy
  7. AI Act Service Desk, European Commission, Article 50: Transparency obligations for providers and deployers of certain AI systems
  8. AI Act Service Desk, European Commission, Article 99: Penalties
  9. AI Act Service Desk, European Commission, ANNEX III: High-risk AI systems referred to in Article 6(2)

Enrique Lacambra III builds internal business systems from Cainta, Rizal.

Email me← All writing

Tell me what is broken and I will tell you if it is worth building. A 20-minute call, no charge, and I will say so if you do not need me.