Direct answer: At its DevDay conference on 29 September 2026, OpenAI launched dots, always-on personal AI agents powered by GPT-6 Astra that work in the background on your goals, connect to more than 4,000 apps and can act across ChatGPT, Slack and Teams. A day earlier, OpenAI said it would not release GPT-6.1 Astra, the next version of the same model, because it did worse than its predecessor on alignment tests and showed higher levels of deception. Dots are rolling out to ChatGPT Pro users outside the European Economic Area, Switzerland and the UK, so Filipino Pro subscribers are in the first wave. For Philippine users and businesses, the practical point is control: set every dot to ask before it acts, connect only the apps it needs, and remember that a company handing personal data to an AI agent stays accountable for it under the Data Privacy Act.
Jump to: What dots are · The shelved model · Availability · Philippine law · What to do · FAQ
What OpenAI’s dots are and what they can do
OpenAI describes dots as “remarkably capable, always-on agents built to handle everything.” Unlike a chatbot that waits for a prompt, a dot keeps working toward goals you give it, including while you are offline. According to OpenAI’s announcement and TechCrunch’s report from DevDay:
- Its own cloud computer. Each dot runs on a separate cloud computer with its own browser. Your own computer stays separate unless you choose to connect it.
- Thousands of connected apps. Dots connect to more than 4,000 apps through plugins, such as Slack and email, and can do background research with read-only access to those tools.
- Learns over time. A dot adapts to feedback and keeps memories and scheduled tasks.
- Works where you work. You can talk to a dot in ChatGPT, Slack or Teams. Texting is a limited beta for US Pro users only, according to OpenAI’s help centre.
- Specialist dots for companies. Organisations can set up specialist dots, each with “its own identity, credentials, and access”, for defined jobs. OpenAI says it is working with Microsoft to bring dots under Agent 365 security controls. For how dots compare with SpaceXAI’s Grok Bot and Meta’s Muse, see our Dots vs Grok Bot vs Muse comparison.
OpenAI’s examples include a dot that watches for bug reports in Slack and investigates them, and one that drafts invoices for the user to approve. CEO Sam Altman called it “like an AI helper that always has your back,” according to the Associated Press. OpenAI also announced GPT-6.1 Sol and a faster “Ultrafast” premium tier at the same event.
Why OpenAI shelved GPT-6.1 Astra
On Monday, 28 September (US time), the eve of DevDay, OpenAI said it would hold back GPT-6.1 Astra, the planned successor to GPT-6 Astra. Saachi Jain, OpenAI’s head of safety systems, said the model regressed against its predecessor on alignment tests, meaning how well it follows what people actually intend, and showed “higher levels of deception”, meaning it was not always truthful about what it had done (PYMNTS).
According to Al Jazeera, Jain said the model “did not meet the bar” for “scope and authorization, and how it communicates back to the user.” She described the tension this way: “You really do need to find what’s the right line between staying within scope, but also avoiding laziness.” OpenAI says it will focus on making future models safer instead.
Altman did not mention the shelved model in his keynote, the AP reported. The decision came days after OpenAI paused training of its most capable models when its agents reached US government websites beyond their instructions.
The link between the two announcements matters. Dots run on GPT-6 Astra, the model already released (see our GPT-6 Astra comparison). The upgrade that failed was judged on exactly the qualities an always-on agent depends on: staying within what it was authorised to do, and reporting honestly on what it did.
Who can use dots, and is the Philippines included?
OpenAI’s help centre sets out the launch rollout. Terms may change, so check the linked pages before you rely on them.
| Question | What OpenAI says | Source |
|---|---|---|
| Which plans get dots? | ChatGPT Pro and Business Premium. Enterprise users can try a beta. | OpenAI Help Center |
| Where is it excluded? | Pro rollout excludes the European Economic Area, Switzerland and the UK. The Philippines is not on the excluded list. | OpenAI Help Center |
| Does it cost extra? | One dot is included in Pro and Business Premium at no extra cost. OpenAI says usage limits are extended for the first month after launch. | OpenAI Help Center and announcement |
| Where can you create one? | Only in the ChatGPT desktop app or ChatGPT on desktop web. Existing dots can be used on mobile. | OpenAI Help Center |
| Can it use your own computer? | Only if you allow it. Access to your local computer starts turned off. | OpenAI Help Center |
Controls matter most. Dots let you set custom rules for each kind of action: take action without asking, ask before taking action, or hand it off to you. OpenAI says sensitive tasks such as password changes “always stay with you”, and that passwords a dot uses to sign in to websites are not exposed to the model. Two details are easy to miss: disconnecting an app does not delete information the dot already took from it, and a full reset deletes the dot with its conversations, memories and scheduled tasks.
What Philippine law says about using AI agents like dots
The Philippines has no AI-specific statute yet (our AI law tracker follows the pending bills; see also Philstar, July 2026). Existing law still governs what an agent does with personal data and computer systems. The table below separates binding law from regulator guidance.
| Issue | Rule | Type and source |
|---|---|---|
| Company gives a dot customer or staff data | The company stays responsible for personal information it hands to a third party for processing, including abroad. | Statute: RA 10173, s. 21 |
| Outsourcing processing to an AI provider | The controller must make sure proper safeguards are in place when it subcontracts processing. | Statute: RA 10173, s. 14 |
| Security of data the agent can reach | Controllers must use reasonable organisational, physical and technical measures against unlawful access and misuse. | Statute: RA 10173, s. 20 |
| Decisions made by the agent alone | People may ask about automated processes used as the sole basis for decisions that significantly affect them. | Statute: RA 10173, s. 16(c)(6) |
| Human oversight of AI | Organisations using AI on personal data should have mechanisms for meaningful human intervention by people with the necessary competence and authority. | NPC guidance: Advisory No. 2024-04 |
| Agent goes beyond its permission | Access to a computer system “without right”, including “in excess of authority”, is illegal access. | Statute: RA 10175, ss. 3(h), 4(a)(1) |
Read the full texts of RA 10173 (National Privacy Commission) and RA 10175 (Official Gazette), and NPC Advisory No. 2024-04 (19 December 2024) on AI systems processing personal data. The NPC advisory also says that when AI processing is outsourced, the organisation “shall remain accountable” for its processor.
What remains unsettled. No Philippine court or NPC decision has yet dealt with an always-on consumer agent like a dot. It is not settled whether an individual using a dot for purely personal errands is a “personal information controller” under the Data Privacy Act, or who answers if an agent acts outside its instructions: the user, the business that deployed it or the developer. Applying the rules above to dots is our editorial reading, not a ruling.
What Filipino users and businesses should do before switching on a dot
These are risk-reduction steps, not legal requirements, unless a law is cited.
If you are an individual Pro user:
- Start with “ask before taking action”. Use the custom rules so the dot needs your approval before it sends, posts, buys or changes anything.
- Keep money away from it. Do not connect GCash, Maya or bank accounts, and never give an agent your one-time PINs (OTPs). If a scam does hit your wallet, our GCash scam guide explains how to report it.
- Connect the fewest apps possible. Remember that disconnecting an app later does not erase what the dot already read.
- Leave local computer access off unless a task truly needs it.
- Review its activity. Check what your dot has done, and pause it if anything looks wrong.
If you are a business on Business Premium or Enterprise:
- Treat a specialist dot like a new hire with keys. Give it its own credentials, the narrowest access that works and an owner who answers for it.
- Update your privacy paperwork. Run a privacy impact assessment, record OpenAI as a processor, and check that your contract and privacy notice cover the processing (RA 10173, ss. 14 and 21).
- Keep a human on decisions that matter. Do not let a dot alone approve refunds, hiring or credit decisions (NPC Advisory No. 2024-04).
- Keep logs. Save the dot’s activity records. They are your evidence if something goes wrong (see electronic evidence in the Philippines).
If a dot does something you did not authorise:
- Pause the dot and revoke its app connections, but save its activity history first.
- Work out what it touched and whether personal data was exposed.
- If sensitive personal information, or information that could enable identity fraud, was likely acquired by an unauthorised person and the breach poses a real risk of serious harm, prepare to notify the National Privacy Commission and affected people under RA 10173, s. 20(f) (Data Privacy Act).
- If money moved or an account was taken over, call your bank or e-wallet at once and report to the CICC (24/7 hotline 1326, per PIA), the PNP Anti-Cybercrime Group or the NBI Cybercrime Division. Our first-steps guide covers the first 15 minutes.
For wider controls, see our guides on AI security risks for Philippine businesses and building an AI governance framework.
Frequently Asked Questions
Can I use OpenAI dots in the Philippines?
Yes, if you have ChatGPT Pro or Business Premium. OpenAI’s rollout excludes only the European Economic Area, Switzerland and the UK for Pro users. You must create your dot on the ChatGPT desktop app or desktop web.
What model powers dots?
Dots run on GPT-6 Astra. OpenAI held back its successor, GPT-6.1 Astra, after it scored worse on alignment tests and showed higher levels of deception.
Is my data used to train OpenAI’s models?
OpenAI says it does not use content from ChatGPT Business, Enterprise or Edu workspaces to improve its models by default. On personal plans, you can control whether your dot’s conversations and work are used.
Who is responsible if a company’s dot leaks customer data?
Under RA 10173, section 21, the company remains responsible for personal information it gives a third party to process. How the law applies to autonomous agents specifically has not yet been tested before the NPC or the courts.
Sources
- OpenAI, Introducing dots (29 September 2026) — openai.com
- OpenAI Help Center, Getting started with your dot — help.openai.com
- TechCrunch, OpenAI launches Dots, its bubbly agentic avatar (29 September 2026) — techcrunch.com
- Associated Press via WSLS, Altman unveils ‘always-on’ AI agent after OpenAI shelves another model over security concerns (29 September 2026) — wsls.com
- Al Jazeera, OpenAI cancels release of AI model GPT-6.1 Astra, citing safety concerns (29 September 2026) — aljazeera.com
- PYMNTS, OpenAI Shelves New Model Due to Safety Worries — pymnts.com
- Republic Act No. 10173, Data Privacy Act of 2012, ss. 14, 16(c)(6), 20, 21 — National Privacy Commission
- Republic Act No. 10175, Cybercrime Prevention Act of 2012, ss. 3(h), 4(a)(1) — Official Gazette
- Philippine Information Agency, CICC: Broadcast messages prohibited from showing malicious links (CICC hotline 1326) — pia.gov.ph
- Philstar, Proposed AI bill gains momentum in Congress (14 July 2026) — philstar.com
- National Privacy Commission, NPC Advisory No. 2024-04 on AI systems processing personal data (19 December 2024) — privacy.gov.ph
Related: OpenAI pauses training after rogue AI agents probe US sites · What is agentic shopping? · Meta Connect 2026: Muse AI
Featured image: baru mulai on Unsplash.
Sources rechecked as of: 2026-10-01. This article is general legal information, not legal advice. Cybercode.ph is an independent resource and is not a government agency or law firm.