Category: AI News

If you were hoping to tell an AI helper, “Book me the cheapest United flight home for Christmas,” and walk away, United Airlines has some news for you: not on their planes.

This week, United said it will block AI assistants from booking flights, changing trips or getting into your United account for you. You can still use AI to research a trip. You just have to click “buy” yourself. Here’s what changed, why United did it, and what it means for your holiday travel.

What United announced

On Wednesday, Nov. 11, United posted a new page explaining its rules for what it calls “third-party AI agents.” An AI agent is an AI tool that doesn’t just answer questions but actually does things for you online, like filling out forms, clicking buttons or making purchases. Think of it as a robot errand-runner with your credit card.

United says it is now restricting those tools on its own website and app, and also through travel agencies and booking sites. Here are the main points:

Why United is doing this

United gave three main reasons:

• Your private info. To book for you, an AI agent may need your MileagePlus login, confirmation numbers, birthday and payment details. United says handing all that over “is a lot of control” and could expose personal information.
• Oops purchases. An AI agent could book the wrong flight, change a trip or spend your miles without you meaning it to.
• Federal rules. Airlines must make sure travelers see certain safety information and answer some questions themselves. An AI clicking “yes” for you could break those rules.

There’s also a business reason. United’s CEO, Scott Kirby, told The Wall Street Journal: “To be honest, I don’t want to create another middleman.” In other words, United wants you dealing with United, not with an AI company standing in between.

United did leave the door open. It says it wants to set up “secure technology integrations” with proper privacy and security protections first, and may work with approved AI tools later.

Is this happening anywhere else?

Yes. United is part of a growing list. Some big banks have put limits on AI agents (we covered that in Big Banks Are Blocking AI Assistants From Your Account), and several large websites have started blocking them too (see Big Websites Are Blocking AI Agents). The Verge reports that eBay also appears to block agents from browsing its site.

Not every airline agrees, though. According to Yahoo Finance, Southwest Airlines launched a ChatGPT add-on in October that lets travelers search its flights and fares. So the rules may be different depending on which airline you fly.

The bottom line

This is less about AI being bad and more about trust. Booking a flight involves money, ID details and safety rules, and United doesn’t want a robot making those calls without a human in the loop. For now, the smart move is simple: let AI do the homework, and you push the “buy” button.

Flying soon? Our step-by-step guide to getting ready for your holiday flight with free AI shows how to use a chatbot for packing lists, TSA rules and refund rights without handing over any account info.

Common questions

Sources

• United Airlines: Unauthorized third-party AI agents or assistants
• The Verge: United won’t let your AI agent book your flight
• The Washington Post: United Airlines blocks AI agents from booking flights
• Engadget: United is restricting AI agents from booking flights
• Yahoo Finance: Bots don’t fly here

Ask ChatGPT for the best slow cooker under $60 and, pretty soon, you might get more than an answer. You might get an invitation to chat with the store itself.

On November 10, OpenAI, the company behind ChatGPT, announced a new feature called Businesses in ChatGPT. It lets companies claim an official profile inside ChatGPT, and some can now add a Business Agent: an AI helper that speaks for that company and answers your questions about its products. It’s rolling out just in time for holiday shopping, so here’s what it is, how to spot it, and how to keep your wallet in charge.

What’s new

OpenAI says about 1.2 billion people use ChatGPT every week, and lots of them ask it what to buy. Now businesses get two new tools:

• Business Hub: Companies can claim their business in ChatGPT, get verified, and update their details. Stores can also upload product lists with prices and whether items are in stock. ChatGPT may use that info when it answers you.
• Business Agents: Once verified, a company can set up its own AI helper. Think of it as a store’s chat assistant, but living inside ChatGPT. It answers questions using information and instructions the business gives it.

For now, Business Agents are limited to certain online stores and online service companies in the U.S. OpenAI calls the whole thing a beta, which is tech-speak for “still being tested.” It says it plans to add more types of businesses and more countries over time.

How it works when you’re shopping

Say you ask ChatGPT to compare two vacuum cleaners. If one of the brands shows up in the answer, you can click its name to see its ChatGPT profile. If that company has a Business Agent, you’ll get the option to start a chat with it. Then you can ask follow-up questions like “What’s the difference between these two models?” or “Can I return it if I bought it on sale?”

A few rules OpenAI has set, in plain English:

• It’s clearly labeled. The Business Agent chat opens separately from your regular ChatGPT conversation, and OpenAI says it’s kept “distinct from ChatGPT’s independent answers.”
• It has to say it’s an AI working for the company. OpenAI’s rules say an agent can’t pretend to be a human, an independent adviser, or something OpenAI endorses.
• It can’t buy things for you. Right now, a Business Agent can explain products and policies and give you a link to the store’s website. You finish any purchase or booking on the store’s own site.
• No pushy sales tricks allowed. OpenAI’s policy bans agents from pressuring you, faking urgency (like “only 2 left, buy now!” when that’s not true), or continuing to push a sale after you say no.
• It shouldn’t ask for sensitive info. The rules say agents can’t ask for things like health details, government ID numbers, or bank and card logins.

Ads are part of this too

Here’s the part worth knowing: businesses can also pay to promote their Business Agent through ChatGPT’s ads. OpenAI says those sponsored spots are clearly labeled and kept separate from ChatGPT’s own answers. So if you see the word “Sponsored,” that’s a company paying for your attention, not ChatGPT’s honest pick. (We covered how ads in ChatGPT work last month.)

What about your privacy?

OpenAI says it does not share your ChatGPT conversations or your Business Agent chats with the business. Companies only see big-picture numbers, like how many times they were mentioned or how many people clicked to their website, and only once enough people have chatted to keep it anonymous.

That said, if you choose to give a business your email or phone number (for example, to get a follow-up), that’s information you’re handing to the company. Only share what you’re comfortable with.

Try this before you buy

Before chatting with any store’s agent, get a neutral starting point. Paste this into the main ChatGPT chat (or any free chatbot):

I'm shopping for [item] and my budget is [amount].
Compare 3 good options from different brands.
For each one, list:
- What it does well
- Common complaints from owners
- What to check about returns and warranty
Don't include any sponsored suggestions. Keep it short and in plain English.

Then, if you open a Business Agent, try asking it the tough questions:

What is your exact return policy for items bought on sale during the holidays?
What does the warranty cover, and for how long?
Are there any extra fees for shipping or returns?

Want more ways to shop safely this season? Learn how to check if an online store is legit with free AI, and see how Walmart lets AI assistants shop for you with spending limits.

Common questions

---

Sources

• OpenAI: Introducing Businesses in ChatGPT (Nov. 10, 2026)
• OpenAI: Business Agents Policy
• OpenAI Help Center: Setting up your business on ChatGPT
• The Verge: ChatGPT is getting verified AI ‘Business Agents’
• Reuters: OpenAI lets businesses claim ChatGPT profiles, adds AI agents for US e-commerce firms
• TechRadar: ChatGPT is adding verified ‘Business Agents’ to help you shop

Back in 2023, Google made a clear promise: if you let its AI chatbot look at your Gmail, Docs and Drive, that content would “not [be] seen by human reviewers.” That promise has quietly changed.

Google added a short update to its old blog posts, dated October 16, 2026. It says that with some newer Gemini features, data from your Google apps “can also be subject to human review in limited circumstances.” Hardly anyone noticed until this week, when people online spotted the edit and the story spread fast. Here’s what changed, who it affects, and the switches you can flip if you’d rather keep a stranger out of your inbox.

What changed

Gemini is Google’s AI chatbot. It can “connect” to your other Google apps, like Gmail, Drive, Calendar, Photos and Google Wallet, so it can answer questions such as “When does my flight leave?” or “Find the plumber’s quote in my email.” Google calls these Connected Apps, and they’re part of a feature it calls Personal Intelligence.

Here’s the shift, using Google’s own words:

• 2023: “If you choose to use the Workspace extensions, your content from Gmail, Docs and Drive is not seen by human reviewers, used by Bard to show you ads or used to train the Bard model.” (Bard was Gemini’s old name.)
• Now: Google’s help page says “a subset of Connected Apps data processed by Gemini is reviewed by human reviewers (including trained service providers).” In plain English, that means some of it may be read by real people, including workers at outside companies Google hires.

The help page also gives a pretty blunt warning: “Please don’t connect your apps to Gemini Apps if they contain confidential information that you wouldn’t want a reviewer to see.”

What reviewers might actually see

Google says reviewers don’t scroll through your whole inbox. When Gemini answers a question, it pulls in summaries, excerpts and guesses (Google calls them “inferences”) from your emails and files. Those bits are what can get reviewed. But Google also admits that when an email or file is short, or is especially relevant, “the summary may just be the email or file.” So a short bill, receipt or note could be shown in full.

Why do humans look at all? Google says reviewers check whether Gemini’s answers were wrong, low quality or harmful, and suggest better ones. That’s used to improve its AI. Google says it takes steps first, such as disconnecting the chat from your Google account and “reducing personal information.”

Two more details from Google’s Gemini privacy page are worth knowing:

• Reviewed chats are kept for up to three years, even if you delete your Gemini history later.
• Your regular Gemini chats are kept for 18 months by default, unless you change the setting.

This covers personal Google accounts using the Gemini app. It also matters most if you’ve already said yes to connecting your apps. If you’ve been tapping “No thanks” on those pop-ups, Gemini isn’t reading your Gmail in the first place.

How to lock it down in about 5 minutes

Menu names can shift a little between the phone app and the website, but here’s what to look for.

1. Disconnect apps you don’t want Gemini reading

In Gemini, open Settings and look for Personal Intelligence or Connected Apps. Turn off anything you don’t want shared, like Gmail, Drive or Photos. You can still use Gemini for general questions without them.

2. Turn off Keep Activity

Go to your Gemini Apps Activity page and turn off Keep Activity. Google says this stops your future chats from being used to improve its AI. Even with it off, Google still keeps chats for 72 hours to run the service and keep it safe.

3. Shorten or clear your history

On that same page, you can delete past chats and change auto-delete from 18 months to 3 months. One catch: anything that’s already been reviewed by a person is kept for up to three years anyway, so deleting it won’t pull it back.

Use a free chatbot to check what’s in your inbox

Not sure what’s sitting in your email that you’d rather keep private? Ask any chatbot for a quick checklist. (No need to paste your actual emails in, which kind of defeats the purpose.)

Make me a simple checklist of the kinds of emails and files that most people would want to keep private, like bank, medical and tax items.
For each one, give me:
- A search word I can type into Gmail to find them
- Whether I should delete, archive or keep it
Keep it short and in plain English.

For a broader tune-up, try our online security checkup with free AI. And if you use Chrome, see what to know about Chrome autofill with Gmail and Gemini.

Common questions

---

Sources

• Google: About personalization with Connected Apps (Gemini Apps Help)
• Google: Gemini Apps Privacy Hub
• Google Blog (Sept. 2023, with Oct. 16, 2026 update): Bard can now connect to your Google apps and services
• Google Blog (Oct. 2023, with Oct. 16, 2026 update): Our responsible approach to building guardrails for generative AI
• AI Weekly: Google Gemini Routes Personal Drive, Gmail to Human Reviewers
• Hacker News discussion

Teens have been fibbing about their birthday online for about as long as there’s been an internet. Instagram is now using AI to catch them, all around the world.

On Monday, November 9, Meta (the company that owns Instagram and Facebook) said it is expanding its AI age-guessing technology to every country where Instagram works. If the AI thinks an account belongs to a teen, it moves that account into Instagram’s safer Teen Account settings, even if the teen typed in an adult birthday when signing up. Here’s how it works, what it looks at, and what to do if it gets your age wrong.

What Meta announced

Meta says the change will place “millions more teens” into Teen Account protections on Instagram. The company was already using this technology in the U.S., the European Union and Brazil. Now it’s going global.

A few other pieces came with the news:

• A heads-up for U.S. parents. This month, parents in the U.S. will start seeing a short video in their Instagram feed that explains how teens get placed in the safer settings and how parents can help make sure their teen’s age is right.
• More apps are next. Meta says it’s still rolling the same kind of technology out on Facebook, and plans to start launching it globally on Threads later this year.
• Apple’s age signal in Australia. In Australia, Meta has started using an age range that Apple’s App Store can share when a parent sets up a teen’s iPhone.
• A nudge to rivals. Meta called on TikTok and YouTube to use the same kind of age checks.

Meta also tied the move to its recent agreement with 52 state and territory attorneys general, from both parties, in which it promised to do a better job finding teens and kids under 13 on its apps.

How does the AI guess someone’s age?

It’s not asking for a driver’s license. Instead, the AI looks for clues, a bit like a detective. According to The Verge, those clues include:

• Who and what the account interacts with, like other teen profiles and the kind of posts it engages with.
• Birthday messages. A comment that says “Happy 15th!” is a pretty big hint.
• Visual clues in photos and videos, such as a person’s height and bone structure.

That last part sounds a little creepy, so it’s worth being clear about what Meta says it is not doing. When it first described the photo checks in May, Meta said: “We want to be clear: this is not facial recognition.” The company says the AI estimates a general age range and does not figure out who the person in the picture is.

What changes for a teen’s account

Teen Accounts are Instagram’s built-in safety settings for users ages 13 to 17. Once an account lands there, it gets:

• Private by default, so strangers can’t see the teen’s posts unless the teen accepts them.
• Limits on messages, so people the teen doesn’t follow or isn’t connected to can’t send them direct messages.
• Stricter content filters on what shows up in their feed and searches.
• Time reminders to help limit how long they spend scrolling.

Teens under 16 can’t loosen these settings without a parent’s or guardian’s OK. Parents can also turn on supervision to set daily time limits and keep an eye on how their teen uses the app.

For kids under 13, which is too young for Instagram under its rules, it’s stricter. Meta says the AI is also used to find those accounts. TechCrunch reported that a suspected underage account is shut off, and the person has to prove their age or the account gets deleted.

What if the AI thinks you’re a teen and you’re not?

No AI is perfect, and some adults are bound to get flagged by mistake (looking young has its downsides after all). If that happens, Instagram will ask you to confirm your age. According to Instagram’s Help Center, you can do that by uploading a government photo ID or by recording a short video selfie that estimates your age.

One tip: only confirm your age through the Instagram app itself. If you get an email or text with a link asking for your ID “to unlock your account,” don’t tap it. That’s a classic scam trick.

Use a free chatbot to plan the “birthday talk”

Telling a teen “I’m checking your Instagram settings” can go about as well as asking them to clean their room. A free chatbot like ChatGPT, Gemini or Copilot can help you plan a calmer conversation. Paste this in and fill in the brackets:

Help me plan a short, friendly talk with my [age]-year-old about their Instagram account.
I want to:
- Check that the birthday on their account is correct
- Explain why Teen Account settings exist, without lecturing
- Agree on a few simple rules together, like time limits and who they can message
Give me 3 easy conversation starters, 2 things to avoid saying, and a fair compromise I could offer if they push back. Keep it warm and under 200 words.

For more help keeping kids safe with AI tools, see our guide to ChatGPT’s teen protections for parents.

Common questions

---

Sources

• Meta: Expanding Teen Account Protections to Teens Around the World
• The Verge: Instagram’s AI age detection is expanding globally
• Mashable: Meta announces global roll out of its age estimation system
• TechCrunch: Meta will use AI to analyze height and bone structure to identify if users are underage
• Instagram Help Center: How video selfie age verification works

ChatGPT just got a playful new trick. On November 4, 2026, OpenAI rolled out Images 2.6, and with it you can make looping animated GIFs from a text prompt or a still photo. No special design app required.

If you have ever wanted a tiny looping clip for a group chat, a party invite or a silly pet reaction, this is the update that makes that feel normal. Here’s what launched, what everyday people can actually do with it, and a few limits worth knowing.

What OpenAI launched

Images 2.6 is live in ChatGPT, Codex and the API for all subscription tiers, on web, desktop and mobile. OpenAI says images now generate about 50% faster than before. People create about 4 billion images a week across ChatGPT, Codex and the API, which is more than half a billion a day.

The headline feature for most of us is motion: ChatGPT can turn words (or a photo you upload) into a looping GIF. GIF creation is still rolling out, so you may not see it in your account on day one.

How GIF creation works in ChatGPT

According to OpenAI’s help docs, you can:

• Ask for a looping GIF and describe what should move. Example: “Make a looping GIF of a paper boat bobbing on the water.”
• Upload a still photo and ask ChatGPT to animate it.
• Adjust playback speed after the GIF is made, and edit frames in the chat.
• Save or share the finished GIF from ChatGPT.

One important limit: ChatGPT cannot edit an animated GIF you upload. It can edit GIFs it created for you, and it can animate a still image you upload. So if you already have a finished GIF from somewhere else, this tool won’t tweak that file.

Saved reference photos for consistent faces

Images 2.6 also makes it easier to keep the same person, pet or character looking like themselves across edits. You can upload a photo and ask ChatGPT to save it as a reference. Later, type @ plus the name you gave it in your prompt. OpenAI’s example: if you saved your dog as Pepper, try “Make a watercolor portrait of @Pepper.”

That means fewer “wait, that doesn’t look like my dog” moments, and you don’t have to re-upload the same photo every time.

A quick note for developers

In the API, OpenAI offers two flavors: GPT-Image-2.6 Flare (faster; as little as about 6 seconds at low quality) and GPT-Image-2.6 Sunburst (higher quality and consistency). Most readers can skip the API details. The everyday takeaway is the same: image tools are getting quicker and more controllable.

Try a simple first prompt

Once GIF creation shows up in your ChatGPT account, paste something like this:

Make a short looping GIF of a steaming coffee mug on a wooden table. Soft morning light, gentle steam rising, cozy and calm. No text, no logos.

Then ask it to slow the playback down, or change one detail (“make the mug blue”). Small edits in chat are part of the point of Images 2.6.

Common questions

---

Sources

• OpenAI: Introducing Images 2.6
• OpenAI Help: Images in ChatGPT
• The Verge: ChatGPT will now make you animated GIFs

If your child or grandchild uses a school Google account, a big change is coming. On Tuesday, November 3, Google said its AI tools will be switched off for K-12 student accounts until the school decides to turn them on. It’s also launching a new “coach” mode meant to help kids learn instead of just handing them answers.

Sounds simple, right? There’s a catch, and it has a lot of teachers and parents talking. Here’s what changed, what the fuss is about, and what it means for your family.

What Google announced

Right now, many schools that use Google’s tools have its AI chatbot, called Gemini, turned on for students. Google says that will flip. Under the new rules, these tools will be off by default for K-12 student accounts:

• Gemini, Google’s AI chatbot (similar to ChatGPT)
• Gemini Notebook, a study tool that works with your own notes and files
• Gemini in Classroom, the AI tab inside Google Classroom, the app many teachers use to hand out and collect homework
• Google Search itself

A school’s tech administrator (the person who manages the school’s accounts) has to switch them on. Education Week reports that schools can choose different settings for different grades, and that Google expects the changes to start early next year. Google says it will share details with schools in the coming months.

The catch: no Search without AI

Here’s the part that surprised people. Google says Search and AI are now “deeply integrated,” so schools can’t turn on a plain, old-fashioned Google Search with no AI summaries. If a school wants AI off, students lose Google Search on their school accounts too.

“If a school is not comfortable with that and simply wants AI off, then they have the choice to not have Search on at all,” Chris Phillips, who runs Google’s education business, said in a meeting about the changes, according to Education Week.

That puts some schools in a tough spot. New York City has banned student-facing AI through 8th grade, and Los Angeles schools paused AI on their devices for a year. Both use Google products. Critics didn’t hold back: Senator Richard Blumenthal accused Google of “prioritizing profits over people,” and a leader at Common Sense Media, a nonprofit that reviews kids’ media, told Education Week that Search without AI would be the best option, but it’s “an option that Google is not providing.”

What is Guided Learning?

Google’s answer to the “kids just copy the answer” worry is a feature called Guided Learning. Instead of solving the problem, it acts like a patient tutor. It asks questions and breaks the work into steps so the student does the thinking.

• In Google Classroom: Rolling out worldwide starting November 3, for schools that already have Gemini turned on. Teachers can tie it to their class materials, and Education Week reports teachers can read the conversations students have with it.
• In Google Search: A test in select U.S. schools starts later this month. If a student pastes in a physics problem or asks for help with a book report, Google says Search will walk them through it step by step instead of jumping to the answer.

Will it work? It’s too early to say. Common Sense Media reported in July that Google’s AI search mode gave full answers to all 180 homework assignments its testers tried. Google says Guided Learning is built to fix exactly that, but outside groups haven’t had a chance to test it much yet.

Turn any free chatbot into a tutor at home

You don’t have to wait for Google. You can ask ChatGPT, Gemini or Copilot, all free, to act like a coach instead of an answer machine. Have your student paste this in before starting homework:

Please act as my tutor. I'm in [grade] and I'm working on [subject or assignment].
- Don't give me the final answer, even if I ask for it.
- Ask me one question at a time to help me figure it out myself.
- If I get stuck, give me a small hint, then let me try again.
- When I finish, check my work and explain any mistakes in simple words.
Here's the problem: [paste the problem or question]

It’s not perfect. A determined kid can always open a new chat. But it sets the tone. For more ideas, see our guide to helping with homework using free AI.

Common questions

---

Sources

• Google: Designing responsible AI for learning
• Education Week: Google’s new rules hand schools hard choices about AI and Search
• AP News: Google will cut off search access for students at schools that reject AI
• Techzei: Guided Learning in Google Classroom and new school AI controls

AI chatbots don’t run on magic. They run on electricity, and lots of it. Every question you type into ChatGPT gets answered by powerful computers sitting in giant buildings called data centers. As more of those buildings go up, a lot of people are asking a fair question: will my electric bill go up to pay for them?

On Monday, November 2, OpenAI (the company behind ChatGPT) made a new promise about that. It’s called the Grid Protectors Program. Here’s what it is, what it isn’t, and what it means for your power bill.

What happened

OpenAI says its future AI buildings, which it now calls “compute campuses,” will be designed to use less electricity when the power grid is under strain. The grid is the big web of power plants, power lines and equipment that brings electricity to your home. It gets stretched thin on the hottest summer afternoons and the coldest winter mornings, when everyone cranks the AC or the heat at once.

• Turning down on busy days. OpenAI says it will work with local power companies to agree on how much electricity its campuses can stop using during those crunch times, how fast and for how long. The details will be different in each place.
• Paying its own way. OpenAI says families and small businesses “should not be asked to subsidize” its electricity needs, and that it will pay for the power lines and equipment its campuses require.
• Backup power. Where it makes sense, OpenAI says it may also use on-site power or big batteries to help smooth things out.

The first test: Georgia

The program’s first real example is a planned OpenAI campus in Effingham County, Georgia, about 45 minutes outside Savannah. It’s still being developed. The deal was made with Georgia Power, the state’s biggest electric company, and approved by state regulators in August.

• It’s huge. Georgia Power says the project adds 3,200 megawatts of new demand. A megawatt is a big chunk of electricity; for comparison, OpenAI says 1,000 megawatts is about what 700,000 homes use on average.
• Up to 1,000 megawatts can be turned down. When demand is high, Georgia Power can cut the power going to the campus by up to that amount. That’s a little under a third of the campus’s total need.
• OpenAI pays for the hookup. Under the deal, OpenAI pays the full cost of the infrastructure needed to serve the site.

Georgia Power says its contracts with big new customers like this one, plus expected growth in the state, should save a typical home using 1,000 kilowatt-hours a month at least $15 a month, or $180 a year, starting in 2029. Think of it like a group dinner: if a new friend with a big appetite joins and pays their full share, everyone else’s part of the bill gets a little smaller.

The fine print

Before you start planning how to spend that $15, a few honest caveats:

• It’s a promise, not a result yet. The Georgia campus isn’t running, so nobody has seen this flexibility work in real life.
• The savings aren’t all from OpenAI. Georgia Power’s estimate covers all of its large new customers, not just this one. And it’s a projection, not a guaranteed credit. Fuel prices and storm repairs can still push bills up.
• It only applies in some places. The $15 figure is for Georgia Power customers. Other areas will have their own deals, or none at all.
• Not everyone is convinced. The Verge noted the announcement comes as opposition to data centers spreads across the U.S., and that several states have passed laws meant to stop AI companies from passing energy costs on to residents.

Let free AI explain your electric bill

Electric bills are full of confusing lines like “delivery charge,” “fuel adjustment” and “kWh.” (A kilowatt-hour, or kWh, is how electricity is measured. Running a 1,000-watt space heater for one hour uses about 1 kWh.) Paste this prompt into ChatGPT, Gemini or Copilot, all free, and type in the numbers from your bill:

I'm not an energy expert. Please explain my electric bill in plain English.
- My power company: [name]
- How much electricity I used this month: [number] kWh
- Total bill: $[amount]
- The charges listed on my bill: [type each line and its amount, for example: Customer charge $10, Delivery $35, Fuel $28]
Tell me what each charge means, which ones I can lower by using less electricity, and 5 simple ways to cut my use this winter. Don't guess at my power company's rates or programs; tell me what to look up on their website instead.

Leave out your account number and address. The AI doesn’t need them to help. And if you’ve got other bills that make your eyes cross, here’s how to get AI to explain any confusing letter or bill.

Common questions

---

Sources

• OpenAI: Introducing the Grid Protectors Program
• The Verge: OpenAI pitches data center power plants as a way to protect nearby areas
• Georgia Power: Contract with OpenAI approved, part of portfolio delivering about $950 million in annual savings
• 41NBC: Georgia Power says OpenAI deal could save customers $15 monthly
• AI Affairs: OpenAI launches grid programme with up to 1,000 megawatts of flexibility in Georgia

If you’ve priced a new laptop lately and done a double take, you’re not imagining things. This week, chipmaker Intel gave a big warning: computers have gotten so much more expensive that people are buying fewer of them, and the company doesn’t expect that to change soon. The reason? Artificial intelligence, or AI.

The short version: the giant computer centers that power AI tools like ChatGPT are buying up huge amounts of memory chips. That leaves fewer chips for everyday laptops and desktops, so prices go up. Here’s what happened, and what it means if a new computer is on your holiday list.

What happened

On Thursday, October 29, Intel shared its quarterly results. Intel makes the main “brain” chip (the processor) inside many Windows laptops and desktops, so it has a good view of the whole computer business. Its CEO, Lip-Bu Tan, told investors: “Our customers began signaling a PC slowdown based on the memory prices and availability.”

• Fewer computers sold. Intel expects worldwide PC sales to drop about 12% this year, with the second half of the year down more than 20%. It’s planning for another drop of about 10% in 2027.
• No quick fix. Intel’s finance chief said the slowdown should last at least through the first half of next year. He also said Intel doesn’t expect the memory supply to improve during 2027, which leaves 2028 as a possible turning point.
• Wall Street noticed. Intel’s stock fell about 29% on Friday, October 30, its worst day since the 1970s. Computer makers like Dell and HP slid too.

Why AI is making computers cost more

Every computer, tablet and phone needs memory chips. The main kind is called RAM (random access memory). It’s your computer’s short-term workspace. Think of it like counter space in a kitchen: the more you have, the more you can work on at once without things piling up.

AI tools run in data centers, which are huge buildings packed with powerful computers. Those machines use a special, pricey kind of memory built for AI. Memory makers earn more selling that kind, so they’ve been shifting their factories toward it. That leaves less of the regular memory that goes into laptops, desktops and phones.

Less supply plus steady demand means higher prices. Research firm IDC expects the average price of a PC to rise 17% in 2026, and it predicts PC sales will fall about 20% in the last three months of this year, normally the busiest shopping season. Back in March, another research firm, TrendForce, estimated that a typical $900 laptop could cost close to 40% more once higher memory and processor costs get passed along.

So if your next laptop feels pricey, you can blame the robots. Or at least the robots’ landlords.

How to shop smart for a computer right now

• Check what you have first. On Windows, go to Start, then Settings, then System, then About, and look for “Installed RAM.” On a Mac, click the Apple menu, then About This Mac, and look for “Memory.” Write it down so you can compare.
• Buy the memory you’ll actually use. Memory is measured in GB (gigabytes). For email, web browsing, video calls and streaming, 8 GB can get by, but many reviewers suggest 16 GB for a Windows laptop you plan to keep for years. Don’t pay extra for 32 GB unless you edit videos or play big games.
• Look at certified refurbished. These are returned or used computers that have been checked, fixed if needed and resold. Apple, Dell and Amazon (through Amazon Renewed) all sell them. Make sure there’s a warranty and a return window.
• Consider a Chromebook for simple jobs. If you mostly use the web, email and video calls, a Chromebook is often cheaper and doesn’t need much memory to run smoothly.
• Watch out for “too good to be true” deals. Scammers love holiday shoppers. Buy from stores and sellers you know, and pay with a credit card for extra protection.

Let free AI compare laptops for you

Spec sheets can look like alphabet soup. Paste this prompt into ChatGPT, Gemini or Copilot (all free) and fill in the brackets:

I'm shopping for a laptop and I'm not a tech person. Please help me in plain English.
- What I use a computer for: [for example: email, Facebook, video calls with family, online banking]
- How long I want to keep it: [for example: 5 years]
- Laptop 1: [paste the name and details from the store page]
- Laptop 2: [paste the name and details from the store page]
Tell me which one fits me better and why. Point out if either one has too little memory (RAM) or storage for what I do. Then give me 3 questions to ask before I buy. Don't guess at prices; tell me to check them myself.

AI can get product details wrong, so double-check anything important (like the warranty) on the store’s own page. For more help planning holiday spending, try our guide to making a holiday gift list and budget with free AI.

Common questions

---

Sources

• CNBC: Amazon, Intel shares fall after earnings amid AI spending concerns
• TechCrunch: Intel’s stock crashes, knocking $12B from government’s stake
• The Motley Fool: Intel Q3 2026 earnings call transcript
• TheStreet (via The State): Intel CEO flags PC memory crunch
• Intel: Third-quarter 2026 financial results
• PCWorld: Holiday PC sales are expected to plunge due to memory shortages
• TrendForce: Rising memory and CPU prices could push notebook prices up by nearly 40%
• Reuters: Smaller phone and laptop makers dig in for years of memory scarcity

If there’s a teenager in your house, there’s a good chance they’ve asked ChatGPT for help with homework, a text to a friend, or something you’d rather not know about. On October 30, OpenAI, the company behind ChatGPT, announced new safety limits for teens, plus a few changes that apply to everyone.

The short version: teens will get fewer answers about weapons and security systems, ChatGPT can now end conversations that keep circling back to violence, and the tools that guess a user’s age are getting stricter. OpenAI says the changes roll out over the next few weeks.

What’s changing

Here’s what OpenAI says is new, in plain English:

• Tighter limits for teens. Teen accounts will get less practical information about weapons and how security systems work, even when nothing in the chat sounds threatening. Teens can still get help calming a situation down, staying safe, or reaching a trusted adult.
• Chats can be ended. For users of any age, ChatGPT can now automatically end a conversation that shows a fixation on violence. If it happens several times, OpenAI may pause the account for a while.
• Less detail about past attacks. ChatGPT will give fewer specifics about past mass attacks, no matter how old the user is.
• Better age guessing. OpenAI uses an age prediction system, which is software that looks at signals like how an account is used to guess whether someone is under 18. It will now keep checking over time and move likely teens into the teen version of ChatGPT. Adults who get put there by mistake can prove their age through a company called Persona.
• A public safety report. OpenAI says it made more than 200 referrals to law enforcement from July through September. Its first Community Safety Report comes out in early 2027, then every six months.

How to link your teen’s account

OpenAI already offers free parental controls. Here’s how to turn them on, according to OpenAI’s help pages:

• On your own ChatGPT account, go to Settings, then Parental controls, then Add family member.
• Choose to send an invite by email or phone number.
• Your teen accepts the invite from the link, or from Settings, then Parental controls, then Accept invite.
• Once you’re linked, you can set quiet hours (times ChatGPT won’t work), turn off voice mode or image making, and keep the extra content protections on.

One thing to know: linking does not let you read your teen’s chats. OpenAI may send you a safety alert if it spots a serious risk, like signs of self-harm. Since July, parents also get a notice if a teen’s account is shut down for breaking the violence rules, and OpenAI says notices for account pauses are coming.

Not sure how to bring it up? Try something like, “I’m not trying to spy on you. I just want us both to know how this thing works.” Teens can smell a lecture from a mile away, so keep it short.

The bigger picture

This comes as AI companies face growing pressure over how chatbots treat kids. Earlier this month, several states sued OpenAI over child safety, which we explained in States Sue OpenAI Over ChatGPT and Kids. Reuters notes that, by OpenAI’s own earlier count, teens spend under 15 minutes a day on ChatGPT on average.

If your kids mostly use AI for schoolwork, our guide to AI homework help for parents shows how to make it a tutor instead of an answer machine.

Common questions

---

Sources

• OpenAI: Community safety report and new protections for teens
• Reuters (via Yahoo News): OpenAI to limit weapons information for teens
• CNA: OpenAI to limit weapons information for teens on ChatGPT
• Bloomberg Law: OpenAI expands ChatGPT protection for teens
• OpenAI Help: Managing parental controls in ChatGPT
• OpenAI Help: ChatGPT for teens

Ever spent ten minutes clicking around Amazon just to skip one Subscribe & Save delivery? Or tried to find out where a refund went, and ended up five menus deep? Amazon wants you to just ask instead.

On October 29, Amazon announced that its AI helper, Alexa for Shopping, can now handle more than 100 new everyday tasks on your account. You type or say what you want in plain words, like “Skip my next delivery” or “Where is my refund?”, and it takes care of it. Amazon says it’s available now to all US customers.

What is Alexa for Shopping?

Despite the name, you don’t need an Echo speaker. Alexa for Shopping is a chat helper built right into Amazon. You find it by tapping the Alexa icon in the Amazon Shopping app on your phone, or at the top of the screen on Amazon’s website on a computer.

It could already help you compare products, check up to a full year of price history, set “Update Me When” alerts for restocks and new releases, and even buy something automatically when it drops to a price you pick. Amazon says more than 400 million customers used it in the last 12 months.

What’s new: 100+ things you can just ask for

Amazon didn’t publish the full list, but it gave plenty of examples. Here are the most useful ones, with the exact phrases Amazon suggests:

• Lists and registries. “Create a new shopping list,” “Add all to my baby registry,” or “Show me what’s left on my wedding registry.” It can also move items from one list to another, which is handy for holiday wish lists.
• Subscribe & Save. “Skip my next delivery,” “Pause my subscription,” or “Change my delivery to every three months.”
• Prime and other memberships. “What are my Prime benefits?”, “How much have I saved with Prime this year?”, or “When does my Audible subscription renew?”
• Delivery instructions. “Change the delivery instructions on my account,” for example to tell drivers to leave packages at the side door.
• Billing and refunds. “I have an unrecognized charge on my account,” “Show my gift card balance,” or “Where is my refund?”

“Alexa for Shopping can now help with more than 100 brand-new shopping actions,” said Rajiv Mehta, Amazon’s vice president of Conversational Shopping, “giving customers the ability to handle everything from building a holiday gift list to resolving a billing question just by asking.”

How to try it

• Update the Amazon Shopping app on your phone, then open it and tap the Alexa icon. On a computer, go to Amazon.com and look for Alexa for Shopping at the top of the page.
• Start with something low-stakes, like “What are my Prime benefits?” or “Create a new shopping list called Thanksgiving.”
• Then try a real chore, like “Show my Subscribe & Save deliveries for next month.” Read what it plans to do before you confirm any change.

Tip: If you spot a charge you don’t recognize, asking Alexa for Shopping is a fine first step. But if you think someone got into your account, change your Amazon password and call your bank or card company too. And remember that Amazon will never ask you to pay with gift cards or share a one-time code over the phone. If someone does, it’s a scam.

The bigger picture

Big stores are racing to put AI helpers in charge of your shopping. Walmart is pushing its own AI shopping agents, and some banks and stores are pushing back on outside AI bots that try to shop or log in for you. Amazon’s move keeps the AI inside its own app, where it can see your orders and account.

Want to read more? See our story on Walmart’s AI shopping agents and why some banks are blocking AI agents. And while you’re cleaning up subscriptions, try our guide to finding forgotten subscriptions with free AI or making a holiday gift list that fits your budget.

Common questions

---

Sources

• Amazon: Alexa for Shopping adds 100+ new capabilities
• The Verge: Alexa for Shopping adds lists and gift registries (report)
• Android Authority: Alexa for Shopping update (report)
• Digital Trends: Alexa for Shopping’s 100 new features (report)