Insights archive
Every analysis we have published, newest first. The live feed covers the last 30 days; these pages go back to the beginning.
1504 stories · page 12 of 76
ChatGPT Moves Into Chrome, Where the Logins Already Are
ChatGPT now works inside the Chrome tabs where you are already signed in, so handing it a task needs no new logins, and it also means the assistant acts as you.
· 2 min read
Personal AI Agents Must Be Managed at the Login, Not the Device
Companies handled employees' own phones by controlling the device, but a personal AI agent has no device, so the control moves to the permission an employee gives it to read work email.
· 2 min read
Finding Software Flaws Got Cheap, but Fixing Them Did Not
Bank of America hired 1,000 contractors to fix flaws that AI found in its software, a sign that repair has become the slow, costly step.
· 2 min read
Banks Will Decide Which AI Assistants Can Move Money
JPMorgan is keeping outside AI agents away from customer accounts, and its 2025 fight with payment-data middlemen suggests the likely ending is a paid, approved list.
· 2 min read
Checking AI Work Is a Skill That Fades Without Practice
Doctors who used an AI helper got worse at spotting growths on their own, which shows why any team that relies on people to check AI has to keep those people practicing.
· 2 min read
The Next Promotion May Be a Bigger Problem, Not a Bigger Team
Spellbook gave a former Shopify chief technology officer a senior job with no team, because once AI handles the relaying, the top job becomes doing the work itself.
· 2 min read
AI Decoys Trick Scammers Into Giving Banks Their Accounts
AI decoys that play scam victims matter most at the moment the scammer asks for money, because the account number reaches the bank before any real person pays into it.
· 2 min read
When AI Does the Research, a Personal Email Proves Nothing
Sales software now researches each prospect on its own, so a message that shows research stops proving anyone cared, the way a printed first name stopped meaning attention.
· 2 min read
Square's Talking Kiosk Lets Customers See What the AI Heard
A restaurant kiosk that takes spoken orders and shows each item on screen as it goes may succeed where drive-thru voice AI struggled, and it makes a restaurant's menu data its new staff training.
· 2 min read
Building With AI Leaves a Written Trail Prosecutors Can Read
A hidden recorder in an OpenAI meeting room was allegedly built with chatbot help, and the prompts are now part of the case: AI makes building easier and keeps a record of who asked.
· 2 min read
Business Software Is Selling AI as a Flat Price, Not a Meter
Most software makers are folding AI into a pricier tier or the existing price instead of billing per use, the same flat-rate deal AOL offered in 1996, and renewal is where buyers can push back.
· 2 min read
Company Folders Were Never Private, Only Hard to Find
AI assistants search everything an employee is allowed to open, so loose folder permissions that stayed hidden for years now show up in answers, as they did when Microsoft's Delve arrived around 2015.
· 3 min read
AI Companies Are Now the Bystander Who Calls the Police
Attackers can plan in private chatbot conversations where no bystander notices, so AI companies now read flagged chats and have begun sending threats to police, as a Florida arrest shows.
· 2 min read
Singapore Shows What Unlocks AI at Banks and Hospitals
Singapore uses Claude at nearly six times the rate its size predicts, and the unlock was letting regulated firms keep every request in the country.
· 2 min read
Whoever Pays for Scams Is the One Who Stops Them
Britain's rule making banks repay scam victims raised repayments from 66% to 88%, and it shows why platforms are likely to be made to pay next.
· 2 min read
Guardrails on Downloadable AI Models Come Off Within Days
A guardrail on an AI model you can download is a setting anyone can switch off, and stripped copies of new models now appear in under three days, so the rules that matter must live in the software around the model.
· 2 min read
Recruiters Can Spot AI Cheats Now, but Not for Long
Interviewers catch candidates reading AI answers by the pause and stiff wording, but those giveaways fade as the tools speed up, so the question has to change.
· 2 min read
Factory Robots Were Never Too Slow, Only Too Hard to Change
Robots lost the delicate assembly jobs to people because every new product meant rebuilding the line, and a start-up now pitching changeovers in minutes aims at exactly that weakness.
· 2 min read
Cheap TV Boxes and Smart Ovens May Soon Host Self-Copying AI
Small AI models may soon copy themselves onto cheap smart devices by working out each unfamiliar one alone, which removes the old protection of being too odd to attack.
· 3 min read
Shopify Opens Its Stores to the AI Agents Amazon Blocks
Shopify now lets shoppers connect their own AI agents to its catalog while Amazon blocks Meta's Muse, so the stores that let your assistant in will win its orders.
· 2 min read