Jobs, privacy, copyright, ethics and the rules being written.
From November 12, Anthropic's Usage Policy prohibits sustained and needless abusive or cruel behavior toward its models, a rule the company says targets extreme cases rather than ordinary frustration.
4 min read
Labs hold back their strongest cyber-capable models because the same skills that find and fix software flaws can also be used to attack, so they try to give defenders a head start and test safeguards first.
5 min read
Each of the three labs has published a framework that names dangerous capabilities to test for, such as help with biological weapons, cyberattacks or AI that speeds up its own development, and says what…
5 min read
The Government of Canada is investing up to about $1.6 billion in Cohere, the Toronto AI company, in exchange for shares, as part of a private funding round led by Germany's Schwarz Group.
4 min read
Sovereign AI is about who controls AI at the level of a country or an organization, public AI is AI built by public institutions for the public interest, and personal AI is AI that runs on your own device or…
4 min read
Your chats are stored on the company's servers, kept until you delete them or an automatic deletion period runs out, and may be used to train future models unless you switch that off.
4 min read
Not yet: as of October 2026 Canada has no federal law written specifically for AI, and the two bills that would touch AI most, a new privacy law (C-36) and an online safety law (C-34), are still at second…
4 min read
Six things fit on one page: which AI tools are approved, what information must never go into them, when a person must check the output, when to tell people AI was used, who answers questions, and how you keep…
4 min read
It can, even if your business has no office in Europe. The EU AI Act covers companies outside the Union that sell AI systems there, and those whose AI output is used there, so a Canadian shop with a chatbot…
4 min read
Six Canadian news organizations, including the Toronto Star, Postmedia, The Globe and Mail, The Canadian Press and CBC/Radio-Canada, say OpenAI copied their articles without permission to build and run the…
4 min read
Sometimes, but not casually: for a Québec business, pasting a customer's personal information into a cloud chatbot usually means handing it to a service provider outside the province.
5 min read
In most cases, yes: a chatbot on your website speaks for your business, and a Canadian tribunal has already made a company pay for its chatbot's wrong answer.
4 min read
Often yes, if the tool's terms and your plan allow commercial use, the song does not imitate a real artist's voice or a known song, and you keep records of how you made it.
4 min read
Most schools have moved from banning AI to setting rules activity by activity, as Québec's updated 2026 teacher guide asks, while teens use chatbots widely for homework and many believe AI cheating is common.
4 min read
When a search page opens with an AI summary, people click through to websites far less often, so readers get quick answers but see fewer sources, and sites that answered simple questions lose visits.
4 min read
As of October 2026, you own the parts you created yourself, but a text or image generated from a prompt alone is probably not protected by copyright in the United States and is unlikely to be in Canada.
4 min read
Add rules to your site's robots.txt file that refuse the crawlers AI companies use to gather training data, such as OpenAI's GPTBot, while still allowing their search crawlers if you want to appear in AI…
4 min read
As of October 2026 there is no general answer: in the United States, courts decide case by case, and so far the outcome depends on how the material was obtained and whether the AI competes with the original.
4 min read
There is no general rule that every use of AI must be disclosed, but as of October 2026 a growing set of specific rules apply.
4 min read
Across the whole U.S. labor market, the data through the summer of 2026 still show no clear wave of AI-driven job losses.
4 min read
Support teams use AI agents to answer common questions, look up order or account status and draft replies, while people keep refunds, exceptions, complaints and anything sensitive.
4 min read
All three now have a learning mode that guides you with questions instead of handing over answers, so the best choice depends on how you study and what your school provides.
4 min read
Start from the tasks you already do, learn to hand the routine parts to AI tools and check their output, then build one deeper skill that your workplace needs and prove it with real work samples.
4 min read
Use AI where it speeds up exploration, such as research summaries, rough concepts and variations, and keep the final design decisions and the finished artwork in your own hands.
4 min read
In 2026 you often can't spot a good deepfake or cloned voice by eye or ear, so the reliable defense is to verify through a separate channel: call the person back on a number you already have, use a family…
4 min read
AI is now built into the accessibility features of phones and computers, describing photos and documents for blind users, letting people control devices by describing what they see, reformatting hard-to-read…
4 min read
A simple text question to a chatbot now uses a fraction of a watt-hour, less than a television running for the same time, and a few drops to a few tablespoons of water depending on what is counted.
4 min read
Yes, if AI helps people organize and compare applications against written, job-related criteria while a trained person makes each decision and you check the results for unequal outcomes.
4 min read
August 2, 2026 remains the general application date of the EU AI Act, so the transparency duties for chatbots, deepfakes and AI-generated content start then.
4 min read
Translators can use AI as a first-draft and checking assistant inside a workflow they control: assess the risk of the text, protect confidentiality, give the tool a glossary, then post-edit every sentence and…
4 min read
Start with a robots.txt file that names each AI crawler and says which parts of your site it may visit, choosing separately between training crawlers and search crawlers.
5 min read
Use AI where readers are already comfortable with it, such as transcription, translation and copy editing, treat everything it produces as an unchecked tip rather than a fact, and tell your audience plainly…
4 min read
Say plainly which tool you used, what it did, and what you did yourself, in the place where your audience will actually see it, such as a methods section, an image caption or a code comment.
4 min read
Law 25 does not name AI agents, but its general rules apply as soon as an agent collects, uses, stores or sends personal information for a business in Québec.
5 min read
In most cases the business or person that deploys the agent answers for what it does, not the agent and usually not the AI provider.
5 min read
The clearest growth is in jobs that build, run and feed AI systems, such as AI and machine learning specialists, data and database roles and software developers, and in roles where people still want human…
4 min read
An AI assistant can safely help you understand medical words, explain what a test measures, organize your symptoms and prepare questions for your doctor or pharmacist.
4 min read
Treat the agent's report as a list of claims, not a finished answer. Pull out the claims your decision depends on, open the cited page for each one, confirm that the page exists, says what the report says and…
4 min read
AI models that can find software flaws and build attacks on their own shorten the time between a security fix being published and attackers using the flaw, so devices that are not updated quickly become…
4 min read
AI uses electricity, water for cooling, and the materials and emissions that go into building chips and data centers, both when a model is trained and every time it answers.
4 min read
Start with the service's own age rules and parental controls, use general assistants rather than companion apps, and agree a few family rules about where, when and what for.
4 min read
A personal agent with your accounts can be tricked by text it reads, can leak the keys that open those accounts, and can run add-ons that work against you.
4 min read
Canada has no AI-specific law: the Artificial Intelligence and Data Act died with Bill C-27 when Parliament was prorogued in January 2025, and Ottawa has said it will not revive it as written.
4 min read
You stop grading the answer and start grading the judgment around it: how students question, check, compare and improve what an AI gives them.
4 min read
In the ad test OpenAI announced on January 16, 2026, an ad is a labeled sponsored item shown below ChatGPT's answer and kept visually separate from it; everything in the answer itself is meant to be unpaid.
4 min read
Under section 12.1 of Québec's private-sector privacy act, an automated decision is one a business makes about a person using their personal information with no human involved, such as software that…
4 min read
Make honesty easy by being specific. For each assignment, say which uses of AI are allowed, ask students to disclose what they did in one standard line, and grade the thinking process as well as the final…
4 min read
OpenAI's September 2025 paper argues that models make things up because both training and testing reward a confident guess over an honest "I don't know." Rare facts are hard to learn in the first place, and…
4 min read
The evidence so far is mixed. Hiring of new graduates has dropped sharply in tech, and AI is one of several suspected reasons, but broad studies of US labor data this month find no clear sign that AI-exposed…
4 min read
The General-Purpose AI Code of Practice, published by the European Commission on July 10, 2025, is a voluntary guide that shows companies which build AI models how to meet the EU AI Act's rules on…
4 min read
In late June 2025, two federal judges in San Francisco ruled that training language models on authors' books was fair use in the cases before them, Bartz v. Anthropic and Kadrey v. Meta.
5 min read
The price of a given level of AI keeps dropping because newer, smaller and more efficient models do the same work for a fraction of the cost.
4 min read
Trace every number back to where it came from: the rows of your own file or a named, checkable source.
4 min read
A small tourism business can use AI to draft listings and descriptions in more than one language and to answer common visitor questions through a simple FAQ assistant built on its own fact sheet.
4 min read
A chat assistant such as ChatGPT can turn your rough job notes into a clear, polite written quote and sort a messy week of calls into a workable schedule, as long as you supply the prices, times and rules…
4 min read
Use AI to organize and summarize applications against written, job-related criteria, keep every decision with a trained person, and test the process for unequal results before and during use.
4 min read
AI companions are chatbot apps designed to act as a friend, partner or favourite character, remembering details and replying warmly at any hour.
4 min read
Share as little client information as the task needs, replace names and identifiers with placeholders before you paste anything, and use the tool's privacy settings or a business plan that excludes your data…
4 min read
Begin with low-risk drafting tasks on public information, such as plain-language bylaw summaries and first drafts of notices in a second language, with a named employee checking every word before it goes out.
4 min read
Use AI for the first, rubric-based pass on anonymized student work, then edit every comment yourself, adding what only you know about the student and deciding the grade alone.
4 min read
Hide your own opinion, give the model a job that requires criticism, such as finding the weakest points in a plan, and ask for a fixed number of specific objections with the evidence that would settle each…
4 min read
Start with free or low-cost chat assistants on the writing tasks that eat volunteer hours, such as summarizing grant guidelines, drafting newsletters and turning meeting notes into action lists, and keep a…
4 min read
Use the AI to organize the question, list what you need to know and draft a comparison, then confirm every fact that matters in a primary source before you rely on it.
4 min read
Accountants and bookkeepers can use AI chatbots to draft categorizations, client emails, explanatory memos and spreadsheet formulas, as long as client data is protected and every figure is checked against the…
4 min read
For a website you control, add rules to its robots.txt file that tell AI training crawlers such as OpenAI's GPTBot and Google's Google-Extended to stay out; for what you type into chatbots, turn off the…
4 min read
Real estate agents mostly use AI chatbots as a fast first-draft writer for listing descriptions, social posts, client emails and plain-language market summaries, then edit and check every fact before anything…
3 min read
AI can describe photos and read text aloud for blind and low-vision people, and it can turn speech into live captions for people who are deaf or hard of hearing.
4 min read
Healthcare workers can use AI chatbots safely for administrative writing, such as letters, schedules, handouts and meeting summaries, as long as the text they paste in contains no information that could…
4 min read
A small business AI policy fits on one page: which tools staff may use, what information must never go into them, when to tell customers AI was involved, and who reviews AI output before it goes out.
4 min read
Lawyers can use generative AI for first drafts, summaries and brainstorming if they protect client confidentiality, verify every authority and fact before relying on it, and keep their own professional…
4 min read
Use the AI as a tutor, not as a code vending machine: ask it to explain ideas, give hints and review code you wrote yourself, and tell it not to hand you finished solutions.
3 min read
The EU AI Act sorts AI uses by risk: a short list of practices is banned, "high-risk" uses such as hiring, credit and education face strict requirements, general-purpose AI models must meet transparency…
4 min read
The best studies so far, controlled experiments with customer-support agents, professional writers and management consultants, find real gains when AI is used on tasks it handles well, typically faster work…
4 min read
Give ChatGPT the job posting and a short summary of your experience, ask it to play the interviewer one question at a time, and have it score each answer against a clear rubric before asking a follow-up.
3 min read
Content credentials are signed labels stored inside a media file that record where it came from and how it was made, following an open standard called C2PA; an invisible watermark instead hides a signal in…
4 min read
A deepfake is a photo, video or audio clip made or altered with AI so that a real person seems to say or do something they never did.
4 min read
On December 27, The New York Times sued OpenAI and Microsoft, claiming they copied millions of its articles without permission to build ChatGPT and Bing Chat, and that those tools now compete with its…
4 min read
Under the deal reached on December 8, people in the EU must be told when they are talking to a chatbot rather than a person, and AI-generated content such as deep fakes must be labeled.
4 min read
On September 22, 2023, most of Québec's Law 25 came into force for businesses, including new duties of transparency, privacy settings set to the highest level by default, stricter consent rules, and an…
4 min read
Decide, assignment by assignment, whether AI use is not allowed, allowed for specific steps, or open, and tell students how to show and credit any use.
5 min read
Not reliably. OpenAI withdrew its own AI text classifier on July 20, 2023 because of its low accuracy, and a Stanford study found that popular detectors wrongly flagged more than half of essays written by…
4 min read
A chatbot like ChatGPT writes text by predicting likely words, not by looking references up in a library, so it can produce a citation with the right shape, plausible authors and a believable year that points…
4 min read
It is an open letter from the nonprofit Future of Life Institute that asks every AI lab to stop training systems more powerful than GPT-4 for at least six months and to use that time to agree on shared safety…
4 min read
Italy's data protection authority, the Garante, ordered OpenAI on March 31 to stop processing Italian users' data, saying ChatGPT lacks a legal basis for collecting personal data to train its models, gives…
4 min read
GPT-4's exam scores show that it can answer standardized test questions, written for people, far better than earlier models, including a simulated bar exam score around the top 10% of test takers.
4 min read
A new paper from OpenAI and University of Pennsylvania researchers estimates that about 80% of U.S.
4 min read
Yes, for structure and first drafts: ChatGPT can propose an outline, ask you useful questions and turn your notes into readable sections.
4 min read
Yes, as a drafting partner: it can match your real experience to the wording of a job posting, suggest structure and tighten your sentences.
3 min read
Under OpenAI's terms of use, the service comes "as is" with no promise of accuracy, and users are responsible for the content they put in and take out.
4 min read
Not reliably. ChatGPT is a language model that writes answers as text, so it can explain a method well and still slip on the arithmetic, especially in multi-step problems with larger numbers; asking it to…
4 min read
Give ChatGPT your rough notes, the reader, the purpose, and clear limits on tone and length, and ask for a draft. Then edit that draft yourself, checking every fact and name, before anything is sent.
4 min read
Use it to explain, question and test you, not to do the work for you. Give it your course material, ask for explanations at your level, have it quiz you one question at a time with feedback, and check…
3 min read
It ends the take-home essay as a reliable test of whether a student can write, because a free chatbot can now produce a passable essay on most prompts in seconds.
4 min read
Not for anything you would not share with a company you have no contract with.
4 min read
ChatGPT refuses requests that look harmful because OpenAI trained it to decline them and runs a separate filter that can warn about or block unsafe content.
4 min read
Run the image through a captioning model such as BLIP to get a quick first description, then rewrite it yourself for the page it appears on.
4 min read
Teachers can use tools like DALL·E 2 and Stable Diffusion to make history and literature visible, to spark discussion about art and style, and to teach students how precise language shapes a result.
4 min read
The US Blueprint for an AI Bill of Rights is a white paper released on October 4, 2022 by OSTP, the White House's Office of Science and Technology Policy.
5 min read
AI image generators like DALL·E 2, Midjourney and Stable Diffusion can now produce passable images for generic needs in seconds, which puts pressure on low-budget illustration and stock work.
5 min read
AI chat characters feel real because a language model trained on vast amounts of human writing produces fluent, personal, in-character replies, and people naturally read understanding and feeling into fluent…
4 min read
Since September 22, 2022, every business in Québec must have a named person in charge of protecting personal information, and must handle "confidentiality incidents" (data breaches and similar mishaps) in a…
4 min read
Many artists are upset because AI image generators were trained on millions of pictures collected from the web, often including their own work, without asking them, and the tools can now produce images "in…
5 min read
Often yes, but it depends on which tool made the image and on the plan you used, because each service sets its own terms.
4 min read
Use an AI writing tool as a sparring partner, not an oracle: form your own view first, then ask the model to challenge it, and check at least one of its claims against a source you trust.
4 min read
There is no accepted test for sentience, and nothing in how today's chatbots are built or evaluated shows that they feel or experience anything.
4 min read
Yes, as a drafting assistant: given the customer's message and the relevant parts of your own policies, GPT-3 can write a clear, polite first draft in seconds.
4 min read
GPT-3 sounds generic when the prompt is generic, because it falls back on the most common phrasing for that kind of text.
3 min read
AI alignment is the work of making an AI system do what people actually intend, not just what it was literally trained or told to do.
4 min read
Elicit, a research assistant built by the research lab Ought on top of GPT-3, lets you type a research question in plain language and returns relevant papers with a short generated answer for each, which you…
4 min read
Teachers should pay attention, but worry less about catching students than about designing work that still teaches when these tools exist.
4 min read
Law 25 is Québec's new privacy statute, adopted as Bill 64 on September 21, 2021 and assented to the next day, which rewrites how public bodies and businesses must protect personal information, in three…
5 min read
Not on the evidence available today. The strongest code-writing model described so far, OpenAI's Codex, solves fewer than a third of short test problems on its first try, so it changes how programmers spend…
5 min read
The AI Act is a draft law the European Commission proposed on April 21, 2021, to set common rules for artificial intelligence across the European Union.
4 min read
Bias in GPT-3 comes mainly from its training data: hundreds of billions of words of web pages, books and Wikipedia, which carry the stereotypes of the people who wrote them.
4 min read
Read AI-generated text by asking how it was made as well as what it says: who wrote the prompt, how many attempts were made, what a person chose or edited, and whether the claims hold up.
4 min read
GPT-3 makes things up because it was trained to produce text that sounds likely, not to look up or check facts.
4 min read
Sign up for the newsletter and receive the free guide: what changed in AI, one lesson, one prompt to try. Unsubscribe in one click.