Connect with us

Technology

OpenAI says its latest GPT-4o model is ‘medium’ risk

Published

on

OpenAI says its latest GPT-4o model is ‘medium’ risk

OpenAI has released its GPT-4o System Card, a research document that outlines the safety measures and risk evaluations the startup conducted before releasing its latest model.

GPT-4o was launched publicly in May of this year. Before its debut, OpenAI used an external group of red teamers, or security experts trying to find weaknesses in a system, to find key risks in the model (which is a fairly standard practice). They examined risks like the possibility that GPT-4o would create unauthorized clones of someone’s voice, erotic and violent content, or chunks of reproduced copyrighted audio. Now, the results are being released.

According to OpenAI’s own framework, the researchers found GPT-4o to be of “medium” risk. The overall risk level was taken from the highest risk rating of four overall categories: cybersecurity, biological threats, persuasion, and model autonomy. All of these were deemed low risk except persuasion, where the researchers found some writing samples from GPT-4o could be better at swaying readers’ opinions than human-written text — although the model’s samples weren’t more persuasive overall.

An OpenAI spokesperson, Lindsay McCallum Rémy, told The Verge that the system card includes preparedness evaluations created by an internal team, alongside external testers listed on OpenAI’s website as Model Evaluation and Threat Research (METR) and Apollo Research, both of which build evaluations for AI systems.

Moreover, the company is releasing a highly capable multimodal model just ahead of a US presidential election. There’s a clear potential risk of the model accidentally spreading misinformation or getting hijacked by malicious actors — even if OpenAI is hoping to highlight that the company is testing real-world scenarios to prevent misuse.

Advertisement

There have been plenty of calls for OpenAI to be more transparent, not just with the model’s training data (is it trained on YouTube?), but with its safety testing. In California, where OpenAI and many other leading AI labs are based, state Sen. Scott Wiener is working to pass a bill to regulate large language models, including restrictions that would hold companies legally accountable if their AI is used in harmful ways. If that bill is passed, OpenAI’s frontier models would have to comply with state-mandated risk assessments before making models available for public use. But the biggest takeaway from the GPT-4o System Card is that, despite the group of external red teamers and testers, a lot of this relies on OpenAI to evaluate itself.

Continue Reading
Click to comment

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Technology

The best deals we’ve found from Amazon’s Big Spring Sale (so far)

Published

on

The best deals we’ve found from Amazon’s Big Spring Sale (so far)

Amazon loves to manufacture an event. March is historically a dry spell for deals; however, with Amazon’s third annual Big Spring Sale, which runs through March 31st, the retail behemoth is hoping to lure in would-be shoppers with the promise of steep(ish) savings and discounts on more seasonal, spring-centric items to hold folks over until Prime Day surfaces at the onset of summer.

The bulk of the deals we’re seeing right now aren’t quite on par with Black Friday or Prime Day, and, as with most shopping events, not everything on sale is worth picking up. That said, Amazon’s latest sale is one of the first big opportunities we’ve seen this year to save — and bypass some tariff-induced pain — especially since some of our favorite gadgets are currently matching their lowest prices to date, including headphones, robot vacuums, and a slew of charging accessories.

To help you sift through it all, we’re focusing squarely on the gadgets that are actually worth picking up, many of which we’ve tested and recommend even at full price. You’re not limited to Amazon, either. Retailers like Best Buy and Walmart are offering the same deals in a lot of cases, meaning you don’t necessarily need to succumb to shopping at the so-called “Everything Store” just to save a buck (or 50).

Headphone and earbud deals

Streaming and soundbar deals

Advertisement

Fitness tracker and smartwatch deals

Other Verge-approved deals

Update, March 28th: Adjusted to reflect current pricing / availability and several new deals, including those for Apple’s latest entry-level iPad, the Breville Barista Express, and JBL’s Flip 7 speaker.

Continue Reading

Technology

Why that $4 charge on your statement could be fraud

Published

on

Why that  charge on your statement could be fraud

NEWYou can now listen to Fox News articles!

A $4 charge is easy to ignore. It might appear as a parking fee, a small purchase or a vague merchant name that doesn’t raise any red flags. Consumer protection groups and law enforcement say scammers are counting on that.

Recent warnings describe a pattern sometimes called “ghost tapping,” where small, unfamiliar charges show up on contactless payment cards without a clear point of purchase. The amounts are easy to miss, but they can signal unauthorized access to your payment method rather than a simple billing error.

Sign up for my FREE CyberGuy Report. Get my best tech tips, urgent security alerts and exclusive deals delivered straight to your inbox. Plus, you’ll get instant access to my Ultimate Scam Survival Guide — free when you join my CYBERGUY.COM newsletter. 

GHOST-TAPPING SCAM TARGETS TAP-TO-PAY USERS

Advertisement

A small, unfamiliar charge can be the first sign your payment details are being tested without your knowledge. (Leonie Asendorpf/picture alliance via Getty Images)

How does ghost tapping work?

Better Business Bureau (BBB) warnings use the term “ghost tapping” to describe these incidents in consumer alerts. In simple terms, it means a payment can be triggered without you actively making a purchase.  

One reported method involves compromised payment terminals that quietly process low-value contactless charges. Another involves unauthorized NFC readers operating at close range, sometimes in crowded places. In these cases, a card or mobile wallet can be charged within inches.

Because these transactions run through standard contactless payment networks, they appear as normal card-present purchases. There is usually no clear sign on your statement that anything was wrong.

The small charges aren’t arbitrary

Low-dollar transactions are used to confirm that a card can be charged without being declined. When a small charge goes through, it signals that the payment details are active and usable. Once that charge goes through, it establishes a working payment path. Scammers can then run additional transactions using the same card details, sometimes across different merchants or terminals. Many people only notice these charges after they post to an account. At that point, the transaction appears as a completed purchase, not a pending authorization.

Advertisement

Why acting quickly matters

This means a charge that looks harmless could actually be the first step in a larger fraud attempt. Over time, these fraudulent charges may extend beyond a single card. If your payment details have been exposed or are stored in a compromised system, they can be reused until you cancel the card. That can result in multiple charges over several days or even billing cycles.

Delays in reporting also affect how quickly you can stop fraud. The Federal Trade Commission (FTC) advises you to report unauthorized charges as soon as they are identified. In practice, that means contacting your card issuer as soon as you see a charge you don’t recognize, either through the issuer’s app or customer service line.

Credit cards typically offer stronger fraud protections than debit cards, where money may be withdrawn directly from your bank account. Under federal law, liability for unauthorized credit card charges is limited, but those protections depend on when you report it.

Scammed? What federal agencies say to do next

If you spot a charge you don’t recognize, acting quickly can limit damage and protect your rights under federal law.

Review transactions as they post

The FTC recommends checking account activity regularly and flagging any unfamiliar charges as soon as they appear.

Advertisement

Contact your card issuer immediately

The Consumer Financial Protection Bureau (CFPB) advises reporting the charge right away and opening a dispute through your issuer’s app or customer service line.

Submit a formal dispute within 60 days

To preserve protections under the Fair Credit Billing Act, you must send a billing error notice within 60 days of the statement that includes the charge.

Understand the investigation timeline

Card issuers must acknowledge your dispute within 30 days and resolve it within two billing cycles, typically no more than 90 days.

Confirm the charge details before disputing

Check the merchant name, date and amount, since some transactions may appear under payment processors or parent companies.

Replace or block the card if needed

If the charge is unauthorized, the issuer can cancel the card and issue a replacement to stop further transactions.

Advertisement

Scammers often use small charges to confirm your card works before making larger unauthorized transactions. (Clara Margais/picture alliance via Getty Images)

Simply keeping an eye on your accounts may not be enough

Even if you check your statements regularly, small charges can still slip through and delay detection.

Timing can work against you

A lot hinges on when you check your account. If you review transactions infrequently or rely on monthly statements, unauthorized charges may only surface after they’ve already posted. By then, multiple transactions can appear across different dates or merchant names, making them harder to spot.

Alerts don’t always catch small charges

Notifications are not always consistent across accounts. Some banks alert you only after a transaction is completed, while others set thresholds that don’t flag low-dollar amounts. That can allow small charges to slip through without notice. If the same payment details are used elsewhere, activity may not appear in one place.

Ways to protect yourself from ghost tapping

These simple steps can help reduce your risk and catch suspicious charges before they turn into bigger fraud.

Advertisement

1) Turn on transaction alerts

Set up real-time alerts for all charges, even small ones, so you can spot suspicious activity immediately.

2) Use mobile wallets when possible

Apple Pay and Google Pay use tokenization, which means your real card number is not shared with merchants.

3) Keep your card secure in crowded places

Contactless cards can be read at close range, so avoid leaving them exposed in busy environments.

4) Check your accounts more frequently

Do not rely on monthly statements. Review transactions every few days to catch issues early.

How broader monitoring can help

Even with these steps, some suspicious activity may not show up in one place or right away. That’s where broader monitoring comes in. Identity and credit monitoring services are designed to track changes tied to your financial and personal information as they happen. This can include transaction monitoring along with three-bureau credit monitoring, so activity linked to your identity can be seen across accounts and credit files at the same time. That includes new inquiries, account openings and changes that would not appear in any one banking app.

Advertisement

Some services also scan known data breach records and dark web marketplaces for exposed personal information. If your data appears in those sources, you can receive an alert tied to the specific identifier that was found.

HOW DEBIT CARD FRAUD CAN HAPPEN WITHOUT USING THE CARD

If suspicious activity is detected, many platforms offer tools to take action quickly. This may include locking your credit file and working with a U.S.-based fraud resolution team to contact card issuers, dispute charges and close affected accounts.

Some identity protection services also include insurance coverage, often up to $1 million per adult, to help cover eligible losses and legal fees, along with 24/7 fraud resolution support to help restore your identity.

No service can prevent every kind of identity theft. If it happens, monitoring and guided support can make recovery easier to manage.

Advertisement

How to check if your personal information was exposed

If you are unsure whether your personal information has already been exposed, take action now. Our No. 1 identity theft protection pick includes a free identity breach scan, which allows you to see whether your data appears in known leaks. Early detection gives you more control and helps you respond before fraud spreads.

See my tips and best picks on Best Identity Theft Protection at Cyberguy.com.

Catching a suspicious charge early and reporting it quickly can help stop further fraud before it escalates.  (M. Scott Brauer/Bloomberg via Getty Images)

Kurt’s key takeaways

A small charge is easy to ignore, especially when it blends in with everyday purchases. But as these reports show, that’s exactly what scammers are counting on. A few dollars can be the first sign that your payment details are active and vulnerable, giving attackers a way to test and expand access without raising alarms. 

The sooner you catch it, the easier it is to stop. This is one of those situations where habits matter. Checking your transactions regularly, acting quickly on anything unfamiliar and adding an extra layer of monitoring can make a real difference. Fraud doesn’t always start big, but it often starts small.

Advertisement

Have you ever spotted a small charge you almost ignored that turned out to be fraud, and how did you handle it? Let us know by writing to us at Cyberguy.com.

CLICK HERE TO DOWNLOAD THE FOX NEWS APP

Sign up for my FREE CyberGuy Report. Get my best tech tips, urgent security alerts and exclusive deals delivered straight to your inbox. Plus, you’ll get instant access to my Ultimate Scam Survival Guide — free when you join my CYBERGUY.COM newsletter. 

Copyright 2026 CyberGuy.com.  All rights reserved.  

Advertisement
Continue Reading

Technology

Suno leans into customization with v5.5

Published

on

Suno leans into customization with v5.5

Suno just released one of its biggest updates yet with v5.5 of its AI music model. Where previous updates focused mostly on improving fidelity and creating more natural vocals, v5.5 is about giving users more control. It includes three new features: Voices, My Taste, and Custom Models.

In the release notes, Suno says that Voices is its most requested feature. It lets users train the vocal model on their own voice. They can upload clean accapellas, finished tracks with backing music, or just sing directly into the mic on their phone or laptop. The cleaner and higher quality the recording, the less data is required. And to prevent someone from simply stealing another person’s voice, Suno requires the user to also speak a verification phrase. Though, this might be possible to fool with existing AI models of celebrity voices.

Once the Voices feature is trained, users can then have an AI version of themselves sing on uploaded music or AI-generated outputs.

To further personalize outputs, Custom Models allows users to train Suno on their own music. Users will need to upload at least six tracks from their catalog and give the custom model a name. Then they’ll be able to use it to guide v5.5 responses to prompts.

Lastly, My Taste learns tastes and preferences over time. What genres, moods, and artists you routinely come back to for prompts, and applies them when using the magic wand to autogenerate styles.

Advertisement

While My Taste will be available to all users, Voices and Custom Models are reserved for Pro and Premier subscribers.

Continue Reading
Advertisement

Trending