ONLINEAGENT_OPS 2026.Q3 HOME HUBSMODELSBUILDINGCOMPANIESFILMMAKINGSECURITYLIVING WITH ITSKILLSALL HUBS, ONE TERMINAL → ARTICLES CRAFT RECORD BLOG TOOLS SKILLS SEARCH /
HOME›ARTICLES›AI THIS WEEK: 5 TO 11 OCTOBER 2026
ARTICLES · TRACKERBEGINNER

AI This Week: 5 to 11 October 2026

A quiet week for AI news, 5 to 11 October 2026: OpenAI’s plan to retire custom GPTs, ads in chat apps, and new guides to AI safety, health and detectors.

READ8 min
WORDS1,679
SECTIONS6
TYPETRACKER
CHECKED12 OCT 26

The week of 5 to 11 October 2026 was quiet for AI news. This site recorded no new model launch that week, and the one switch-off scheduled in it, Google’s old Antigravity coding agent on 5 October, is still scheduled, not confirmed. What did arrive was 40 new plain-English pages on this site, most of them about using AI safely in everyday life. This page covers the few facts that changed, what is due next, and the new pages worth your time.

TL;DR — THE SHORT VERSION
  • A quiet week: no new AI model launch was recorded on this site between 5 and 11 October.
  • Scheduled, not confirmed: Google’s old Antigravity agent for developers was due to switch off on 5 October.
  • Custom GPTs: OpenAI says it plans to retire them. For affected business (Enterprise) workspaces the planned date is 11 December 2026.
  • Ads: Anthropic says Claude will stay ad-free. ChatGPT can show ads on its Free and Go plans.
  • New guides: is ChatGPT safe, do AI detectors work, health questions, mental health, and nine how-tos for everyday jobs.
◈ IN PLAIN TERMS

Training means a company uses your chats as examples to improve its next AI model. A temporary chat is a chat that is not saved to your history. An AI detector is a tool that guesses whether a piece of writing was made by AI. A switch-off (or shutdown) is when a company turns an AI model off, so programs that use it stop working. Scheduled means the company said it would happen on a date; confirmed would mean it said afterwards that it did.

The news: a quiet week

The plain point: no big launch, but two things are worth knowing if you use ChatGPT or Claude.

OpenAI plans to retire custom GPTs. A custom GPT is a version of ChatGPT set up for one job. OpenAI’s help centre says it is “planning to retire” them, and that for affected Enterprise workspaces the retirement is planned for 11 December 2026. That is a plan, not something OpenAI has confirmed as done. Why it matters: if your team relies on a custom GPT at work, check OpenAI’s help page for what happens to it. Details: what GPT means.1

Ads: where you will and will not see them. This week the site added Anthropic’s promise, made in February 2026, that Claude will remain ad-free. OpenAI says ads may appear on ChatGPT’s Free and Go plans, but not on its Plus, Pro or work plans. Google says your Gemini chats are not used to show you ads. Why it matters: if ads in a chat app bother you, this is one real difference between the three. See ChatGPT vs Claude vs Gemini.234

For developers: the “temperature” setting is going away on Claude too. Temperature is a setting that makes an AI’s answers more or less varied. Google deprecated it in the Gemini API on 21 July.5 Anthropic’s developer docs say it, and two similar settings, are not supported on Claude 4.7 and later models, and a request that changes them gets an error. Why it matters: code that sets temperature can fail on newer models. If you only use the apps, ignore this. See Gemini temperature setting deprecated.6

Is ChatGPT safe? The settings that matter

The plain point: ChatGPT is fine for everyday use if you treat it like a capable stranger: useful, but not private and not always right.

Your chats can be used for training unless you switch it off. On the website: open your account menu, then Settings, then Data controls, and turn off “Improve the model for everyone”. OpenAI says your new chats are then not used to train its models. Why it matters: it is one switch, and it is on until you change it. Step by step, with the catches: is ChatGPT safe to use?7

Temporary chats are the nearest thing to a private window. They stay out of your history and are not used for training, but OpenAI may keep a copy for up to 30 days for safety. So they are fine for a one-off personal question, and still not a place for passwords.8

TAKEAWAY

Switch training off once, and never type anything you would not want a stranger to read.

AI detectors: a guess, not proof

The plain point: an AI detector score is not evidence that someone used AI.

Detectors wrongly flag real people. In a 2023 study, seven detectors wrongly marked more than half of a set of essays by non-native English writers as AI-written. Turnitin, whose detector many schools use, says its score should not be the only reason to act against a student, and OpenAI withdrew its own detector in 2023 for low accuracy. Why it matters: if you, or your child, are accused on a score alone, ask for the evidence and show your drafts. See do AI detectors work?91011

Health questions and mental health

The plain point: a chatbot can help you understand and prepare. It cannot diagnose you, and it is not a therapist.

◈ EMERGENCIES FIRST

If a life could be at risk, do not ask a chatbot. Call your local emergency number: 911 in the US, 999 in the UK. In the US you can also call or text 988, the Suicide & Crisis Lifeline; in the UK and Ireland, Samaritans answer free on 116 123.12131415 More numbers are on the mental health guide.

Asking AI about your health. In a large Oxford study published in February 2026, people who used chatbots to work out what was wrong in doctor-written scenarios did no better than people using their usual sources. In the US, health privacy law usually does not cover what you type into an app you signed up for yourself. Why it matters: use a chatbot to explain words and prepare questions for your doctor, not to decide. See asking AI about your health.1617

Talking to a chatbot about how you feel. The American Psychological Association advises against relying on chatbots for therapy or treatment, and says their handling of a crisis is “limited and unpredictable”. Why it matters: if you use one, use it alongside people, not instead of them. See AI chatbots and mental health.18

TAKEAWAY

Use AI to understand and prepare. Leave diagnosis, treatment and emergencies to people.

Scheduled this week, and what comes next

The plain point: most of these are developer model names. If you only use AI apps, the custom GPT date is the one to note.

WhenWhatRead more
5 OctScheduled, not confirmed: Google’s old Antigravity coding agent for developers shuts down; Google names a newer version to use instead.Page: what is switching off
13 OctDue: Ray-Ban Meta Audio glasses ship; Ray-Ban Display glasses reach France, Italy and Germany.Page: Meta Connect 2026
OctoberPromised: the wider launch of Kling 4.0, a video model. Still in early access, with no launch date, on 12 October. Update 3 Nov 2026: Kling says 4.0 “rolled out to all users on October 15”.19Page: the Kling cheat sheet
23 OctScheduled: twelve older OpenAI models shut down, including GPT-4 Turbo and o3-mini.Page: what is switching off
2 NovScheduled: xAI retires one Grok Imagine image model name. Old requests do not simply fail: the newer model answers them at its lowest quality setting.Page: what is switching off
11 DecPlanned: OpenAI retires custom GPTs for affected Enterprise workspaces.Page: what GPT means

20

◈ STILL NOT CONFIRMED

The Antigravity switch-off on 5 October. This site’s last check of Google’s page was on 1 October, before the date; no confirmation has been recorded since. The Sora API, scheduled to shut down on 24 September: OpenAI still had not confirmed it when this site re-checked on 12 October, and it names no replacement (see the Sora API shutdown).21 Update 13 Oct 2026: OpenAI has since confirmed the Sora 2 models and Videos API were shut down on 24 September 2026.22 Claude Haiku 5.5 was promised “in the coming weeks”, with no date, when the Sonnet 5.5 page was checked on 1 October.23 This page reports what the vendors scheduled, not what was tested.

Last week: AI This Week: 28 September to 4 October 2026.

New on the site this week

The plain point: 40 new pages: 21 guides, how-tos and explainers, and 19 AI words in plain English.

Staying safe. Is ChatGPT safe to use?, do AI detectors work?, asking AI about your health, AI chatbots and mental health and is AI dangerous?, which sorts the risks by how sure anyone can be.

Guides for your life. AI for older adults, with a family code word against copied-voice scam calls; AI for people with disabilities; using AI to budget; and AI skills for work. The US Federal Trade Commission warns that a scammer needs only a short clip of a loved one’s voice to copy it.24

How-tos for everyday jobs. A resume, a job interview, proofreading, Excel formulas, transcribing a recording, planning a trip, a meal plan, learning a language and citing ChatGPT in APA, MLA and Chicago.

How AI works. How does ChatGPT work? follows the few seconds after you press send. Types of AI explains the two ways people sort it: by how capable it is, and by the job it does. Pros and cons of AI puts both sides next to each other. Two older guides, Prompting 101 and what makes a good prompt, now start with everyday chat examples instead of image prompts.

AI words in plain English. 19 new pages, from GPT, knowledge cutoff and sycophancy (when an AI tells you what you want to hear) to voice cloning, explainable AI and the Turing test. The full list is on the AI glossary.

The whole of September is on AI This Month: September 2026. New to all this? Start with AI explained for complete beginners.

SOURCES
  1. 1OpenAI Help Center, GPTs in ChatGPT, read at source 5 Oct 2026“GPTs (also called custom GPTs) are versions of ChatGPT configured for a specific purpose.” … “We’re planning to retire custom GPTs” … “For affected Enterprise workspaces, retirement is planned for Dec 11, 2026.” ↑
  2. 2Anthropic, Claude is a space to think, 4 Feb 2026, read at source 5 Oct 2026“Claude will remain ad-free. Our users won’t see ‘sponsored’ links adjacent to their conversations with Claude”. ↑
  3. 3OpenAI, Ads in ChatGPT, read at source 22 Sep 2026 (as cited on this site’s linked page): “Ads may appear for users on the Free and Go plans. Plus, Pro, Business, Enterprise, and Edu accounts will not have ads.”↑
  4. 4Google, Gemini Apps Privacy Hub, last updated 24 Sep 2026, read at source 5 Oct 2026“Your Gemini Apps chats are not being used to show you ads. If this changes, we will clearly communicate it to you.” ↑
  5. 5Google, Gemini API release notes, entry dated 21 July 2026, read in a browser 16 Sep 2026 (the docs refuse automated fetches)“Deprecated parameters: The sampling parameters temperature, top_p and top_k are now deprecated.” ↑
  6. 6Anthropic, Using the Messages API, undated documentation, read at source 5 Oct 2026“The temperature, top_p, and top_k sampling parameters are not supported on Claude 4.7 and later models and Claude Mythos Preview. Setting them to a non-default value returns a 400 error. Omit them from request payloads and use prompting to guide the model’s behavior instead.” ↑
  7. 7OpenAI Help Center, Data controls in ChatGPT, “Updated: 7 days ago”, read at source 5 Oct 2026 in a browser“When Improve the model for everyone is off, your new conversations won’t be used to train OpenAI models. They can still appear in your chat history.” ↑
  8. 8OpenAI Help Center, Data controls in ChatGPT, read at source 5 Oct 2026 in a browser, listing what temporary chats do while they remain temporary, ending “may be retained for up to 30 days for safety purposes”.↑
  9. 9Liang et al., Patterns 4(7), July 2023, read at source 5 Oct 2026the detectors “incorrectly labeled more than half of the TOEFL essays as ‘AI-generated’”. ↑
  10. 10Turnitin Guides, Using the AI Writing Report, read at source 5 Oct 2026 in a browser“it should not be used as the sole basis for adverse actions against a student.” ↑
  11. 11OpenAI, New AI classifier for indicating AI-written text, 31 Jan 2023, with update of 20 Jul 2023, read at source 5 Oct 2026“the AI classifier is no longer available due to its low rate of accuracy”. ↑
  12. 12911.gov, Calling 911, last updated 24 Jan 2023, read at source 5 Oct 2026“In an emergency, dial 911 from your phone immediately.” ↑
  13. 13NHS, Where to get urgent help for mental health, page last reviewed 26 Apr 2023, read at source 5 Oct 2026“Call 999 or go to A&E now if:” … “you do not feel you can keep yourself or someone else safe”. ↑
  14. 14988 Suicide & Crisis Lifeline, 988lifeline.org, read at source 5 Oct 2026“If you need emotional support, reach out to the national mental health hotline: 988.” ↑
  15. 15Samaritans, Contact a Samaritan, read at source 5 Oct 2026“Whatever you’re going through, call us free any time, from any phone, on 116 123.” ↑
  16. 16Bean et al., Nature Medicine, Reliability of LLMs as medical assistants for the general public, 9 Feb 2026, read at source 5 Oct 2026“participants using the same LLMs identified relevant conditions in fewer than 34.5% of cases and disposition in fewer than 44.2%, both no better than the control group.” ↑
  17. 17US Department of Health and Human Services, Protecting the Privacy and Security of Your Health Information When Using Your Personal Cell Phone or Tablet, content last reviewed 29 Jun 2022, read at source 5 Oct 2026“In most cases, unless the app is provided to you by a covered entity or its business associate, the HIPAA Rules also do not protect the privacy of data you’ve downloaded or entered into mobile apps for your personal use”. ↑
  18. 18American Psychological Association, Health advisory, November 2025, read at source 5 Oct 2026“Do not rely on GenAI chatbots and wellness apps to deliver psychotherapy or psychological treatment” and “The ability of these tools to consistently and safely manage a user in crisis is limited and unpredictable.” ↑
  19. 19Kling AI, Kling 4.0 at Work, 27 Oct 2026, read at source 3 Nov 2026“Since Kling 4.0 rolled out to all users on October 15, creative teams across film, advertising, and animation have been putting these capabilities to work.” ↑
  20. 20Google, Gemini API deprecations, last updated 30 Sep 2026, read at source 1 Oct 2026antigravity-preview-05-2026, released “May 19, 2026”, shutdown “October 5, 2026”, replacement antigravity-preview-09-2026 · Meta, Introducing Ray-Ban Meta Audio and More AI Glasses Styles, 23 Sep 2026, read at source 1 Oct 2026: “they ship October 13” · Meta, New Features for Meta Ray-Ban Display, 23 Sep 2026, read at source 1 Oct 2026: “ahead of availability on October 13” · Kling AI, Inside Kling 4.0: A Complete Guide to AI Video Creation, 30 Sep 2026, read at source 12 Oct 2026: “The full model is coming in October 2026.” · OpenAI, Deprecations, read at source 1 Oct 2026: 2026-04-22 entry, “Access to these models will be shut down on October 23, 2026.” · xAI, grok-imagine-image-quality migration guide, read at source 1 Oct 2026: “Requests to grok-imagine-image-quality are served by grok-imagine-image-2.0 with quality set to low”; retirement date “November 2, 2026” · OpenAI Help Center, GPTs in ChatGPT, read at source 5 Oct 2026: “For affected Enterprise workspaces, retirement is planned for Dec 11, 2026.” ↑
  21. 21OpenAI, Deprecations, read at source 12 Oct 2026removal “from the API on September 24, 2026”, with no recommended replacement listed. ↑
  22. 22OpenAI, Sora 2 model page, read at source 13 Oct 2026“The Sora 2 models and Videos API were shut down on September 24, 2026 and are no longer available.”; “There is no one-to-one replacement API.” ↑
  23. 23Anthropic, Claude Sonnet 5.5, read at source 1 Oct 2026“Claude Haiku 5.5 will join the Claude 5.5 family in the coming weeks.” ↑
  24. 24Federal Trade Commission, Scammers use AI to enhance their family emergency schemes, 20 Mar 2023, read at source 5 Oct 2026“All he needs is a short audio clip of your family member’s voice — which he could get from content posted online — and a voice-cloning program.” ↑

OpenAI: GPTs in ChatGPT, Data controls in ChatGPT and New AI classifier, read 5 Oct 2026; Ads in ChatGPT, read 22 Sep 2026; Deprecations, read 1 Oct and 12 Oct 2026.

Anthropic: Claude is a space to think and Using the Messages API, read 5 Oct 2026; Claude Sonnet 5.5, read 1 Oct 2026.

Google, Meta, xAI and Kling: Gemini Apps Privacy Hub, read 5 Oct 2026; Gemini API release notes, read 16 Sep 2026; Gemini API deprecations, Meta’s Ray-Ban Meta Audio and Ray-Ban Display posts and xAI’s migration guide, read 1 Oct 2026; Kling AI, Inside Kling 4.0, read 12 Oct 2026.

Research, health and safety: Liang et al., Patterns, July 2023; Turnitin, Using the AI Writing Report; Bean et al., Nature Medicine, February 2026; US Department of Health and Human Services, health information on your phone; American Psychological Association, Health advisory; 911.gov; NHS; 988 Lifeline; Samaritans; Federal Trade Commission, family emergency scams; all read 5 Oct 2026.

Every source here was read by this site on the date shown and is cited the same way on the linked detail page. The “why it matters” lines are this site’s own plain reading, not claims from these sources.

Free printable AI cheat sheets, and AI news every week.

Sixteen one-page AI guides and checklists in one free PDF, no sign-up. And if you want to keep up, the newsletter covers what changed in AI each week, in plain English, with the sources.

SUBSCRIBE FREE ↗ GET THE CHEAT SHEET PACK
aipromptgeneer://$ about this siteAI PROMPT GENEERAI explained in plain English. Every claim links to its source, with the date it was checked.NEWSLETTERWhat changed in AI each week, in plain English.SUBSCRIBE FREE ↗AI PROMPT GENEER · MACHINE ASSISTANCE DISCLOSEDEVERY ARTICLE CARRIES ITS OWN CHECKED DATE