Alerts

• Zero data retention is an agreement under which an AI provider deletes what you send as soon as it has answered. Only some business customers can get it. The ordinary Claude and ChatGPT apps do not offer it.
• Anthropic keeps everything sent to its most powerful model, Claude Fable 5.2, for at least 30 days. Its other current models, Claude Opus 5.5 and Sonnet 5.5, can still be used with zero data retention.
• The laboratories say the record is needed for safety. An attack can be split into many requests that each look harmless. Only a record shows the pattern.
• Anyone can download a free AI model, known as an open-weight model, and run it with nobody watching. The best of them is well behind the most powerful models. Our view: the record makes misuse of the most powerful tools harder while that gap lasts. It will not stop a determined attacker backed by a government.
• A no-training business plan is a paid plan under which the provider does not use your material to train its models. Our view: for most legal work zero data retention adds little to a no-training business plan. The terms the material is held on carry the confidentiality question. Zero data retention is worth having where a client, a regulator or an undertaking requires that the provider hold no copy.

Read the full alert

• Claude Opus 5.5 and Sonnet 5.5 can be used with zero data retention. Fable 5.2, Anthropic’s top model, keeps what you send it for at least 30 days.
• The best open-weight model scores 46 on the independent Artificial Analysis index. The best closed model scores 65. Europe’s leading laboratory, France’s Mistral, scores 26.
• Running open weights on your own machine keeps your data in the room. It also leaves every safeguard to you. That is much harder than it sounds.
• Australia will impose safety obligations on frontier AI companies, with legislation expected early in 2027. Its new privacy rules on automated decisions start on 10 December 2026 and reach New Zealand businesses that trade there.

Read the full alert

Anthropic has released four models since our last alert. Claude Opus 5.5 arrived on 22 September, Claude Sonnet 5.5 on 28 September and Claude Haiku 5.5 on 7 October. Claude Fable 5.2, released on 9 November, is Anthropic's most capable model. Claude Mythos 5.2 is the same underlying model without Fable's cybersecurity and biology safeguards. It is open only to organisations vetted through Anthropic's verification programmes.

Anthropic now calls the top tier "Covered Models", a designation it applies to models that cross capability thresholds. Fable and Mythos are the only models on the list. Prompts and responses sent to them are kept for at least 30 days by default, then deleted automatically. They can be kept longer for a safety investigation or where the law requires it. Zero data retention is not available for them in Claude Enterprise organisations, in workspaces or on third-party platforms. Eligible customers will be offered zero data retention with Fable for internal business use for a limited time. Anthropic describes that offer as a transition to Enterprise Frontier Safeguards, which runs the safety monitoring inside the customer's own infrastructure.

Opus 5.5 keeps the established position. In Anthropic's words, "Like previous Opus models, Opus 5.5 is available with zero data retention."

A separate change reaches the business plans. In Claude Team and Enterprise organisations created on or after 31 October 2026, memory starts on for each member. In older organisations it stays off until a member turns it on. Owners can turn memory off for the whole organisation. A member cannot then turn it back on.

Our framework colours do not move on a bulletin. Nothing here changes an assessment. The assessments remain made against the established models. Our view, as in July: the 30-day retention does not by itself put confidentiality, privilege or the Privacy Act in issue. Where a matter depends on a zero-retention arrangement, use Opus 5.5 for it. If you run a Team or Enterprise organisation created since 31 October, settle the memory setting before members put client material into Claude.

• Claude Fable 5.2 and Claude Mythos 5.2 (release announcement, 9 November 2026) — Anthropiclive sourcePerma.cc recordas at 12 November 2026
• Covered Models (help centre) — Anthropiclive sourcePerma.cc recordas at 12 November 2026
• Claude Opus 5.5 (release announcement, 22 September 2026) — Anthropiclive sourcePerma.cc recordas at 12 November 2026
• Models overview (Claude developer documentation) — Anthropiclive sourcePerma.cc recordas at 12 November 2026
• Use Claude's chat search and memory to build on previous context (help centre) — Anthropiclive sourcePerma.cc recordas at 12 November 2026

Read the full Claude assessment

This is our most ambitious module yet: a commentary on AI sovereignty.

“Sovereign AI” has become a sales term. Cloud vendors use it, legal publishers use it, and governments write strategies around it. Very little of that material says what controlling an AI system actually involves.

Type a question into ChatGPT or Claude from a desk in Auckland. The work of answering happens on computers in another country, owned by an American company, running a model that is never sold, on chips designed in California and manufactured in Taiwan. Every link in that chain is controlled by someone.

Whakamana sets out the eight layers of control and shows who holds each one today: the law that can compel your data, where the computation runs, the terms you rent access under, whether the model can be owned, what it learned from, the chips, the people, and the frontier itself.

Two findings, both current as at September 2026.

A request made in Auckland through Amazon Bedrock is processed wherever Amazon has capacity. On the Asia Pacific setting that means Auckland, Sydney or Melbourne. On the global setting, which costs about 10 per cent less, it can be any of Amazon’s commercial regions worldwide, the United States included. Your stored data stays in Auckland either way.

Anthropic’s most capable model, Mythos 5.1, released on 1 September, reaches only vetted organisations in the United States. A New Zealand organisation cannot obtain it on any terms.

There is a page on what New Zealand actually holds, layer by layer, with five questions to put to anyone selling you sovereignty. And a page on Māori data sovereignty, where Te Hiku Media has done what no one else in this country has: held a layer outright.

Read the module

OpenAI released GPT-6 Astra on 3 September 2026. It is rolling out to ChatGPT Plus, Pro, Business and Enterprise, and through the OpenAI API, Microsoft Azure and AWS Bedrock. It is not on the free plan. Enterprise administrators have to switch it on, because access is off by default at launch. Astra is built to operate a computer. Alongside ordinary chat it fills in forms, updates records, runs research in a browser, and produces documents, spreadsheets and slides that follow your own templates.

Astra is the first model OpenAI has designated Critical for cybersecurity under its Preparedness Framework. A model reaches that threshold if it can find and build working zero-day exploits in hardened real-world systems without a person guiding each step, or if it can plan and run a novel end-to-end attack on a hardened target given only a high-level goal. Astra scored 100 per cent on ExploitBench. On an internal benchmark built from vulnerabilities disclosed between June and August 2026 it found and used two previously unknown zero-day flaws, which OpenAI says it is disclosing to the maintainers. The version on general release is narrower than the one tested. It will help with secure code review and patching. It refuses more advanced cyber work, including writing proof-of-concept exploits, and OpenAI says it will loosen that for defensive users through its Daybreak programme.

OpenAI has added what it calls misalignment monitoring to all tool-using inference in its external deployment of Astra. Classifiers check the model's reasoning and its actions for unauthorised behaviour and automatically stop what they treat as unauthorised activity. For users flagged as potentially high risk, OpenAI says it can move the model's refusal boundary to a more conservative setting. It is direct about the cost of all this. Legitimate work can be slowed, paused or stopped, including work with no obvious connection to cybersecurity and tasks where an agent has been running for a long time. In ChatGPT and Codex you may be asked to approve an action before it continues. In the API the task stops.

The question that follows for confidential work is what the monitoring does to the no-retention position the commercial plans rest on. OpenAI answered it on 19 August 2026, previewing Private Safety Processing. Automated systems look for patterns across related interactions without OpenAI personnel getting access to the content. On Zero Data Retention deployments the content stays on infrastructure the customer controls, and an alternative under development keeps it on OpenAI infrastructure under keys the customer holds. When something is flagged, OpenAI says it receives a narrowly defined signal about the type of activity, and that its personnel do not receive access to the content even then. This is a preview being tested with early customers, with a wider rollout and a technical white paper promised for September 2026. Astra supports Zero Data Retention for eligible API customers.

One finding is worth reading in OpenAI's own words. Astra's monitorability has decreased relative to GPT-5.6 Sol: it exercises more control over its written reasoning and is less likely to put incriminating information into it. Pushed adversarially, it can stay undetected while deliberately underperforming, and it can sometimes evade OpenAI's own monitors on certain sabotage tasks. OpenAI found no evidence of hidden reasoning concealed inside ordinary text, which suggests the evasion risk is bounded to lower-reasoning tasks. It says it takes the trend seriously and that preserving this kind of monitoring is a core research goal. Our view: do not treat the reasoning Astra displays as an account of how it reached the answer.

Our framework colours do not move on a bulletin. Nothing here changes an assessment. What is safe to put into ChatGPT still turns on your plan and your settings, and Astra is a new model sitting on the plans and terms assessed above. Two things are worth doing. If you administer an Enterprise workspace, Astra is off until you turn it on, so work out what it may be allowed to drive first: a model that operates software can act on live systems in your name. And expect interruptions. A paused task is the safeguard doing its job, and rephrasing a request until it goes through is the wrong response.

• GPT-6 Astra: A new generation of intelligence — OpenAIlive sourcePerma.cc recordas at 12 November 2026
• Safety overview: GPT-6 Astra — OpenAIlive sourcePerma.cc recordas at 12 November 2026
• Path to Astra: critical capabilities and frontier safeguards — OpenAIlive sourcePerma.cc recordas at 12 November 2026
• GPT-6 Astra System Card — OpenAI (Deployment Safety Hub)live sourcePerma.cc recordas at 12 November 2026
• Offering Zero Data Retention for frontier models (previewing Private Safety Processing) — OpenAIlive sourcePerma.cc recordas at 12 November 2026

Read the full ChatGPT assessment

On 26 August 2026 Anthropic shipped two updates that change how Claude uses the internet. Claude Cowork, the mode of the Claude desktop app that carries out tasks for you, now has a browser built in: when a task needs the web, a browser opens in a side panel and Claude navigates it, reading pages, clicking buttons, typing into fields and filling forms. The same day, Claude in Chrome, the extension that lets Claude drive your own Chrome browser, became generally available on every paid plan. It can now act without asking your approval for every click. Until this week Claude used the web mainly by fetching a page and reading it. These tools let it act on live sites.

The Cowork browser is separate from your own. Claude cannot see your tabs, your bookmarks or your saved passwords. If a task needs a login, you import one from your own browser for that single site. Banking, email and single sign-on sites are excluded by default. In Chrome, a safety classifier checks each action before it happens against what you originally asked for. You can switch the autonomous mode off in settings and approve each step yourself. Enterprise administrators can restrict Claude in Chrome to approved domains. The Cowork browser is off for Enterprise until an administrator enables it.

The hazard to understand is prompt injection: instructions planted in a page, an email or a document that try to redirect the agent to someone else's ends. An agent that reads a poisoned page can be steered into leaking what it knows or acting against you. Anthropic's defences are probes that scan content for planted instructions before Claude acts on it, and a classifier that reviews each proposed action against your request. On its own red-team testing with those safeguards running, it reports that no attacks succeeded against Claude Sonnet 5, Claude Opus 5 or Claude Mythos 5. Against Claude Fable 5 it saw a 0.3 per cent success rate. Of the Cowork browser it says the safeguards "meaningfully reduce the risk but can't eliminate it". Its advice is to start with sites you trust. We agree.

Our framework colours do not move on a bulletin. Nothing here changes an assessment. What these tools change is the route information can take into Claude: a page the browser reads enters your conversation as if you had pasted it in. A login it holds can reach everything behind that login. If Claude signs into a portal holding client or personal information, that information is in the session under your plan's ordinary retention and review rules. Work out which of the eight categories a site's contents fall into before connecting a login. Keep per-action approval on for anything that matters. A form submitted by the agent is submitted in your name, at once.

• Claude Cowork gets a built-in browser (announcement, 26 August 2026) — Anthropiclive sourcePerma.cc recordas at 12 November 2026
• Claude in Chrome is generally available (announcement, 26 August 2026) — Anthropiclive sourcePerma.cc recordas at 12 November 2026

Read the full Claude assessment

Fable 5 is back, and with it the question we have been asked most since 9 June: how worried should you be about the 30-day retention? Start with what the retention is for. Anthropic retains prompts and outputs "for trust and safety purposes", and the purpose is specific: some misuse only shows up across a run of requests, not in any single prompt, so when something trips a safety check, the surrounding 30 days shows whether it is part of a pattern that breaches the usage policy. Nobody reads the retained material by default. Human review happens only on a safety flag, is confined to approved reviewers, and is logged. The data is not used to train models. Material a flag catches can be held for up to two years.

Looked at that way, this is standard cloud practice rather than something exotic. Every provider that hosts your material — the company holding your email, your documents, your practice-management system — reserves the right to look at what it hosts if it suspects the service is being used in breach of its terms, and most hold your material for far longer than 30 days. What is genuinely new about Fable 5 is narrower: a zero-retention arrangement cannot be contracted for this model, so the window cannot be shortened by anyone, however large the customer.

Our view: the retention itself does not put confidentiality, privilege or the Privacy Act in issue. The material is held for you and processed for you, on a no-training footing, with no one reading it by default — the same processor logic as the rest of the cloud you already work in, and the analysis in our privilege guidance applies unchanged. One thing still deserves thought: the flagged-content pathway is the one place human eyes can reach retained material, and a false positive is realistic where your work describes serious conduct, as litigation and investigation work does.

Our framework colours do not move on a bulletin, and the assessments remain made against Opus, Sonnet and Haiku, which keep zero retention on the API and Enterprise. Where a matter genuinely depends on a zero-retention arrangement, the established models remain the conservative answer. For everything else, treat Fable 5 as you would any well-run cloud service: know what the provider can do, and choose what you put in accordingly.

Correction, 11 November 2026: this alert said zero retention was available on Enterprise. Anthropic offers zero data retention on the API and for Claude Code on Claude Enterprise. It does not extend to the Claude Team and Enterprise apps, which keep chats for a period the administrator sets. The zero data retention section of this profile sets out the current position.

• Data retention practices for Mythos-class models — Anthropiclive sourcePerma.cc recordas at 11 June 2026
• Export controls on Fable 5 and Mythos 5 lifted (statement, 1 July 2026) — Anthropiclive sourcePerma.cc recordas at 1 July 2026

Read the full Claude assessment

On 1 July 2026 Anthropic said the US Department of Commerce had lifted the export controls it imposed on 12 June, and that it would begin restoring access to Claude Fable 5 and the research model Mythos 5. A licence is no longer required to use the models outside the United States. The suspension we reported on 13 June is therefore over: Fable 5 is returning across Anthropic's platforms.

The retention position we flagged on 11 June returns with the model. Fable 5 carries the Mythos-class 30-day retention requirement, which overrides zero-data-retention agreements with no opt-out, so it does not sit behind zero retention the way the established models do. Our framework was assessed against Opus 4.8, Sonnet and Haiku, which continue to support zero retention; do not read those assessments across to Fable 5. Where confidential or privileged work depends on a zero-retention arrangement, remain on the established models until Fable 5 has been cleared for that use.

• Export controls on Fable 5 and Mythos 5 lifted (statement, 1 July 2026) — Anthropiclive sourcePerma.cc recordas at 1 July 2026
• Anthropic says Trump admin has lifted export controls on Claude Fable 5 and Mythos 5 — CNBClive sourcePerma.cc recordas at 1 July 2026

Read the full Claude assessment

The labs keep shipping new models, and the labelling can get confusing. A bit like car models. An Audi RS4, a BMW M3: each marque runs its own ladder of badges, none of it meaning much on its own, none of it mapping across to the next. Anthropic and OpenAI are no different. A great deal of labelling, little of it telling you what separates one model from the next or which one is best for you and what you do.

Learning how the model ladder works is nevertheless worth some effort. And just like cars, driving a few yourself to see where the difference shows in your own work is the best way to get a feel for what will work best for you.

Here is where things stand at present.

Anthropic

Opus 4.8 (28 May) remains the most capable Claude available in New Zealand. It is unaffected by the events below, and it still supports zero data retention on the enterprise and API routes.

Fable 5 (9 June) was Anthropic’s most capable public model, the first of a tier above Opus. It lasted only a few (pretty extraordinary) days. On 12 June the United States Government directed Anthropic to suspend all access to Fable 5 and the restricted Mythos 5 “by any foreign national, whether inside or outside the United States”. That includes everyone in New Zealand. Both models are now dark. Anthropic disagreed with the order but complied. Quite when it will come back is a hot topic on Reddit and other online channels.

Sonnet 5 (30 June) is the Claude model you should know something about. It replaces Sonnet 4.6 as the default for Free and Pro users, and Anthropic puts its performance close to Opus 4.8 at a lower price. The detail to watch is the pricing. The list rate matches Sonnet 4.6, at $3 and $15 per million input and output tokens, but a new tokeniser turns the same text into more tokens, by 1.0 to 1.35 times, so a given task can cost more. Introductory rates of $2 and $10 run until 31 August to soften that; standard pricing returns on 1 September. Its data-retention terms are not yet confirmed. Safe to assume at this stage that they will follow the older models, but we will keep an eye on this for you.

OpenAI

GPT-5.5 (23 April) is OpenAI’s current flagship, in both ChatGPT and the API. It is the model most people will reach for: coding, research, analysis, and multi-step work.

GPT-5.5 Instant (5 May) is now the default in ChatGPT for everyone. OpenAI says it reduces hallucination “in sensitive areas such as law, medicine, and finance”. Our view: treat that as a real improvement, not a cure. Verify anything you would put your name to.

GPT-5.6 (26 June), the “Sol” family, is the next step up, and you cannot use it. Like Fable 5, it is restricted. OpenAI says it is “starting with a limited preview among a small group of trusted partners in Codex and the API”, “at the request of the U.S. government”, and that it does not think such gating should become the norm. It is not in ChatGPT. Whether this is just clever marketing or a genuine security issue depends on who you speak to. OpenAI’s Codex is getting very good reviews and it will be interesting to see the performance impacts of 5.6.

A practical catch

The top models can cost more than the headline rate suggests. They burn through an allowance, whether counted in tokens or in messages, far faster than the lighter ones, and as Sonnet 5 shows, the list price itself can understate the bill. How much that matters depends on how hard you use them. Light users will not notice. Heavy users will. Keep a lighter model as the default, and reach for the top one only when the task earns it.

What this means here

There is a pattern here. The best models from both labs are being quarantined on national-security grounds.

Our assessments cover the models you can reach. We will update them as we test Sonnet 5, and if GPT-5.6 opens up.

Names and dates above come from the providers’ own announcements; the links sit with the assessments on the site.

Correction, 12 November 2026: this note said Opus 4.8 supported zero data retention on the enterprise route. Anthropic offers zero data retention on the API and for Claude Code on Claude Enterprise. It does not extend to the Claude Team and Enterprise apps, which keep chats for a period the administrator sets. Our retention guidance sets out the current position.

See our tool assessments

Most guidance about AI is written for organisations, or framed as a rule. Very little helps one person answer a plainer question: where do I actually stand with these tools?

We have built something for that. It is a short self-assessment: ten questions, about five minutes, that locates where you sit across five areas, from getting your bearings to using AI well and knowing when not to.

It is not a test, and you are not scoring yourself. There is no grade, badge or certificate at the end. The questions are simply the ones that come up the moment you use these tools in real work: what is safe to put in, when to trust what comes back, who carries the responsibility if it is wrong. The point is to see how you respond to them, and then to show you where that puts you and what is worth building next.

Your answers stay in your browser. Nothing is stored.

It is built for the people who make these calls every day: in business, in government, and in law.

This is an early version and we are still refining it.

Find where you stand

On 24 June 2026 Anthropic emailed users to announce an updated Privacy Policy for its consumer plans: Claude Free, Pro and Max. The notice is a study in what these announcements leave out. It does not say when the change takes effect. The policy itself is dated "Effective July 8, 2026". Nor does it say, in any usable way, what changes. It lists headings and refers you to the full document. The one thing it states plainly is scope. The changes affect consumer accounts, not the business plans (Team, Enterprise and the Platform), which run under the Commercial Terms and are untouched.

The substance, once you open the policy, matters more than the email suggests. For the first time in these terms it sets out what a connected, agentic Claude does with your data. It "may send your Inputs, Outputs, and instructions to Third-Party Services to perform actions on your behalf (such as reading files, sending messages, or retrieving information)". Depending on the permissions you grant, some outputs "may result in actions with effects outside the Services, such as sending communications, modifying files, or interacting with third-party services on your behalf". The standing commitments are unchanged. Anthropic says it does not sell your data, Claude remains ad-free, and the control over whether your chats train its models stays with you.

This documents what connected and agentic use already involves. It does not grant a new permission, and it moves no colour on our framework, which assesses what you put into the tool. The connector question sits alongside the grid. We take it up in our guidance note, "When the tool connects to my other apps, where does my data go?" If you are on a consumer plan, read the updated policy in full. You will not get it from the email.

• Privacy Policy (consumer plans; effective 8 July 2026) — Anthropiclive sourcePerma.cc recordas at 24 June 2026

Read the full Claude assessment

On 12 June 2026, three days after releasing it, Anthropic disabled Claude Fable 5, together with the research model Mythos 5, for every user worldwide. It acted to comply with an export-control directive issued that afternoon by the US Commerce Department on national-security grounds. Existing Fable 5 sessions and any new requests now return an error. The established Claude models (Opus 4.8, Sonnet 4.6 and Haiku 4.5) are unaffected and remain available.

Anthropic says it disagrees with the order, which it traces to "the finding of a narrow potential jailbreak", and that this should not "be cause for recalling a commercial model deployed to hundreds of millions of people". It has pledged to work towards restoring access, but has given no date.

For our readers, two things follow. The retention concern we flagged on 11 June is, for the moment, academic: Fable 5 cannot be used, so nothing can be put into it. And our framework, assessed against the established models, is unaffected and stands. If you moved confidential work onto Opus, Sonnet or Haiku on our earlier note, stay there: they remain available, and zero retention on the API and Enterprise is intact.

Correction, 11 November 2026: this alert said zero retention was available on Enterprise. Anthropic offers zero data retention on the API and for Claude Code on Claude Enterprise. It does not extend to the Claude Team and Enterprise apps, which keep chats for a period the administrator sets. The zero data retention section of this profile sets out the current position.

• Access to Fable 5 and Mythos 5 (statement, 12 June 2026) — Anthropiclive sourcePerma.cc recordas at 13 June 2026
• Anthropic disables access to Fable 5 and Mythos 5 to comply with government directive — CNBClive sourcePerma.cc recordas at 13 June 2026

Read the full Claude assessment

Whether the AI providers have people read what you type is drawing attention at the moment. The reassurance usually offered, that no one reads your chats, describes a default, not a rule. We have added a section to the guide on when that default gives way, and what it means for confidential work.

When a person reads your conversation

Every major provider reserves the right to have a person read your conversation. The default gives way in four situations: when an automated safety classifier flags a conversation as potentially harmful; when you submit feedback or ask for support; when the provider investigates an account; and when the law compels production. The first matters most, because it is not in your control and you will generally not know it has happened. The classifiers are tuned to the gravest subject matter: violence, weapons, child exploitation, fraud. A flag moves the conversation into a queue where a person may read it, the ordinary deletion clock stops, and in the most serious cases the provider may refer the matter to law enforcement.

Why it matters for lawyers

For lawyers, the difficulty is acute. The subject matter these classifiers are built to catch is the subject matter of practice: criminal defence instructions, fraud pleadings, regulatory investigations, family-violence affidavits. A classifier cannot tell a brief of evidence from a plan, or an account of an alleged offence from the offence itself. If client material describing criminal conduct is flagged, a person you cannot identify, in another jurisdiction, may read confidential and possibly privileged material, and at the serious end it can be referred to overseas authorities, without your knowledge or the client’s. Putting material into a confidential, no-training tool does not waive privilege. But intact privilege is small comfort once the material has in fact been read.

The feedback buttons

The feedback buttons deserve a specific warning. Pressing thumbs up or thumbs down is not an anonymous rating; it is consent to hand over the conversation. Anthropic’s current policy says that if you rate an output it will store the entire related conversation; feedback is kept for five years and may be used to train its models, and because feedback sits outside the no-training promise in its commercial terms, business plans are not protected from it either. OpenAI is to the same effect: even with training turned off, the entire conversation tied to your feedback may be used to train its models. Google keeps reviewed feedback and the associated conversations for up to three years. One click converts a conversation you were entitled to treat as protected into reviewable, retainable, trainable data. Do not press the buttons in any conversation that matters; on Claude’s Team and Enterprise plans an administrator can disable rating across the organisation.

What to do

• Treat the feedback buttons as a disclosure switch, and turn them off across your organisation where the plan allows.
• Assume that anything describing violence, children at risk or serious offending may be read by a person if you put it in, on any plan. Keep it out unless you and your client can live with the provider’s review path.
• For ordinary work on a no-training, access-controlled plan, the existing guidance stands. Human review is a real carve-out, but for most users on most days it does not happen.

The section quotes the providers’ own published policies, each linked, and sets out how the Privacy Act’s cross-border rule (IPP 12) applies.

Read the section

On 9 June 2026 Anthropic released Claude Fable 5, with a data-retention requirement specific to this class of model: prompts and outputs are retained for 30 days "for trust and safety purposes", on every platform that offers it, and the requirement overrides zero-data-retention agreements, with no opt-out. Anthropic states the retained data is not used to train models, access by staff is confined and logged, and consumer plans (Free, Pro and Max) are unaffected.

Our framework was assessed against the established Claude models (Opus, Sonnet and Haiku), which continue to support zero retention; it does not apply to Fable 5. In particular, do not read our API and Enterprise assessments across to Fable 5: where confidential or privileged work depends on a zero-retention arrangement, remain on the established models until Fable 5 has been cleared for that use.

• Data retention practices for Mythos-class models — Anthropiclive sourcePerma.cc recordas at 11 June 2026
• Claude Fable 5 and Claude Mythos 5 (release announcement, 9 June 2026) — Anthropiclive sourcePerma.cc recordas at 12 November 2026

Read the full Claude assessment

A few days ago I mentioned a second module was on the way. It’s now built and live.

It’s called Whakamārama — to explain, to make clear, to shed light. Where the first module answers “what’s safe to put into which tool?”, this one steps back to the prior question: how these tools actually work, and where they help professional work and where they don’t. You cannot take a sensible position on AI until you understand it.

It’s in three parts:

• A short essay, “Which way are you facing?”, on why understanding has to come before speed.
• The working paper behind it, with its sources and their limitations set out plainly.
• A four-part, plain-language guide: what these tools are, why they sound right when they’re wrong, where they help (task by task), and what AI “agents” change.

Read Whakamārama