Pete Recommends – Weekly highlights on cyber security issues, August 9, 2025

Subject: OpenAI removes ChatGPT self-doxing option
Source: The Register
https://www.theregister.com/2025/08/01/openai_removes_chatgpt_selfdoxing_option/

OpenAI has removed the option to make ChatGPT interactions indexable by search engines to prevent users from unwittingly exposing sensitive information.The feature rollback follows reports of ChatGPT conversations being discoverable in search results, an option recently extended to ChatGPT users.

Dane Stuckey, CISO of OpenAI, announced the change in a social media post. He described it as a short-lived experiment to help people discover useful conversations.

“Ultimately we think this feature introduced too many opportunities for folks to accidentally share things they didn’t intend to, so we’re removing the option,” he said. “We’re also working to remove indexed content from the relevant search engines. This change is rolling out to all users through tomorrow morning.”

Despite explicit warnings “not to share any sensitive content,” ChatGPT users did so anyway, undermining their own privacy. Similar to the search advertising industry, AI vendors argue that models can do more when they have access to our data and applications, at least in the context of chatbots and agents. But giving AI models access to personal info magnifies the privacy and security risks.

OpenAI’s search scrubbing effort appears to be underway but incomplete. Google Search with the site: operator for chatgpt.com/share no longer returns a list of shared, indexed chats. Bing Search returned thousands of results. DuckDuckGo also returned many. So did Brave Search. We saw personal information in many of these results.

[…]

Tagged:

Filed: https://www.theregister.com/software/ai_ml/


Subject: Silent Push CEO talks cybercrime takedowns with The Register
Source: The Register
https://www.theregister.com/2025/08/03/silent_push_ceo_talks_cybercrime/

interview It started out small: One US financial services company wanted to stop unknown crooks from spoofing their trading app, tricking customers into giving the digital thieves their login credentials and account information, thus allowing them to drain their accounts.”Once we poked at it, we realized, it’s certainly not just them,” Silent Push CEO Ken Bagnall told The Register. “We discovered many, many more branches of crime and money laundering, and that initial thing turned into 1.4 million live hosting sites at any one time. It’s a huge crime organization.”

Silent Push, a cybersecurity takedown firm that works with international law enforcement agencies to take down cybercrime groups, first started tracking this crime organization in 2021.

It would eventually turn out to be Funnull, a Philippines-based company that provides computer infrastructure for hundreds of financial scams.

“Funnull is linked to the majority of virtual currency investment scam websites reported to the FBI,” the Treasury Department said at the time, adding that Funnull’s content-deliver-network-hosted websites cost US victims alone more than $200 million in losses, with an average loss of $150,000 per individual.

Bagnall’s company maps criminal groups as they build out their infrastructure, which allows Silent Push to then monitor all the technical connections between the IP addresses and various domains. The threat analysts collect massive amounts of data on the organizations, which law enforcement can then use to build criminal cases and disrupt their operations.

Tagged:

Filed: https://www.theregister.com/security/cyber_crime/


Subject: States Have More Data About You Than the Feds Do. Trump Wants to See It
Source: New York Times
https://www.bespacific.com/states-have-more-data-about-you-than-the-feds-do-trump-wants-to-see-it/

The New York Times – no paywall – “Critics fear that personal data might be used to monitor immigrants and political foes, and to spread false tales of fraud. As the Trump administration has sought to amass personally sensitive data on millions of individuals in America, it has run into one roadblock. The states, and not the federal government, hold many of the details Washington officials would now like to see. The states administer many safety-net programs funded with federal dollars.

The Trump administration is now expanding its data push to this trove, reaching into domains long controlled by the states — and further into their residents’ lives. This week, 20 states with Democratic attorneys general, along with the District of Columbia and Gov. Andy Beshear of Kentucky, sued the U.S. Department of Agriculture over its demand for data on anyone who has applied for or received food stamps in the last five years.
Abstracted from beSpacific
Copyright © 2025 beSpacific, All rights reserved.

Subject: Foreign adversaries are trying to weaponize open-source software, report finds
Source: Nextgov/FCW
https://www.nextgov.com/cybersecurity/2025/08/foreign-adversaries-are-trying-weaponize-open-source-software-report-finds/407190/

Hacking units affiliated with nation-state adversaries are subtly contributing to open-source software tools and working to insert backdoors into publicly available code used by millions worldwide, new research says.Chinese, Russian and North Korean-affiliated hackers are covertly working to insert backdoor hijacks and exploits into major publicly-available software used by countless organizations, developers and governments around the world, according to findings released Monday by Strider Technologies.

The malicious insertions into these open-source tools could allow hackers to pilfer troves of sensitive data from governments and private sector firms, according to Strider, which analyzed open-source code contributors who have direct affiliations with foreign adversaries.

This week, seven teams will compete at the DEF CON hacker conference, where the Defense Advanced Research Projects Agency will evaluate their AI-powered systems designed to autonomously identify and patch vulnerabilities in open-source code.

Tagged:


Subject: Should Lyft and Uber charge more if your battery is low? California may soon ban that
Source: Route Fifty
https://calmatters.org/politics/2025/08/california-surveillance-pricing-ban/

California lawmakers are backing a bill to ban rideshare services from using customers’ information like a cellphone’s battery life, model and geolocation data to set different prices.Should a ridehailing company such as Uber or Lyft be able to charge you more because its artificial intelligence programming thinks you’re desperate since it knows your phone is about to die?

Not if Hayward Democratic Sen. Aisha Wahab has her way. Her Senate Bill 259 would prevent retailers from using artificial intelligence to jack up prices using the information stored on customers’ phones. That could include the phone’s battery life, whether it’s an older model, what apps are installed, what time of day it is, where its user is located and where they live

“Our devices are being weaponized against us in order for large corporations to increase profits, and it has to stop,” Wahab told the Assembly Judiciary Committee last month.

[but] “Suggestions that our systems manipulate pricing unfairly or discriminate are simply false and not supported by evidence,” wrote Zahid Arab, a spokesperson for Uber, in an emailed statement. Shadawn Reddick-Smith, a representative for Lyft, said in an emailed statement that the company “does not base fares on battery percentage.”


Subject: New research shows Iran’s expansive cyber offensive during ‘12-Day War’ with Israel
Source: Nextgov/FCW
https://www.nextgov.com/cybersecurity/2025/08/new-research-shows-irans-expansive-cyber-offensive-during-12-day-war-israel/407207/

One state-backed hacking group created conflict-themed websites to lure pro-Israel visitors and siphon their data, according to SecurityScorecard. Within hours of June’s 12-day war between Iran and Israel erupting, Iranian state-backed hackers and proxy groups launched phishing campaigns, defaced websites and claimed to have leaked troves of stolen data tied to the conflict, according to new threat intelligence released Tuesday.

Telegram also served as a central hub for recruitment, propaganda and orchestration of cyberattacks, according to some 250,000 messages exchanged by 178 Iranian proxy and hacktivist groups throughout the war that were analyzed by SecurityScorecard’s STRIKE threat intelligence team.

The analysis, one of the first comprehensive overviews of the cyberwarfare aspects of the nearly two-week-long conflict, found that Iranian operations were launched in an effort to intimidate civilians, undermine Israeli morale and amplify Iran’s wartime narrative.

The responding cyber campaigns involved three distinct layers of Iranian‑linked actors. At the ground level, loosely organized hacktivists waged symbolic website defacements and claimed to have leaked data under the guise of pro-Palestinian narratives. Above them, proxies aligned with the Islamic Revolutionary Guard Corps combined ideological motives with precise targeting, and frequently conducted joint hacking operations alongside Lebanese or Afghan cyber brigades.

One of the most active players was Imperial Kitten, a group widely tied to the Islamic Revolutionary Guard Corps. The unit stood up conflict‑themed phishing domains, such as nowsupportisrael[.]com and supportisraelfunding[.]com — deliberately named in a way to imply Israeli backing. Once those domains were created, Iran’s cyber warriors deployed advanced remote access malware onto the websites to harvest data from pro-Israel visitors.


Subject: Home Depot and Lowe’s Share Data From Hundreds of AI Cameras With Cops
Source: 404media.co
https://www.404media.co/home-depot-and-lowes-share-data-from-hundreds-of-ai-cameras-with-cops/ [thx beSpacific]
Hundreds of AI-powered automated license plate reading cameras paid for by Lowe’s and Home Depot and stationed in the hardware stores’ parking lots are being fed into a massive surveillance system that law enforcement can access, according to records obtained using a public records request.
The records, obtained from the Johnson County, Texas Sheriff’s Office by the Electronic Frontier Foundation (EFF) and shared with 404 Media, show the sheriff’s office is able to tap into Flock license plate reading cameras at 173 different Lowe’s locations around the U.S. and that it can tap into cameras and gunshot-detecting microphones at dozens of Home Depot stores within Texas. The records are the latest to shed light on how expansive Flock’s surveillance network has become, and highlights that it includes cameras that are operated by both police and private businesses.[…]
Flock’s automated license plate reader (ALPR) cameras are stationed along roads or at entrances to parking lots around the United States, and constantly scan the license plates of cars that drive by. Because there are Flock cameras around the country, Flock often has a snapshot of people’s movements which police can search, typically without a warrant.
[…] more Flock articles: https://www.404media.co/tag/flock/

Subject: Uber Gets Report of Sexual Misconduct Every 8 Minutes
Source: Newser [+ AI]
https://www.newser.com/story/373093/uber-gets-report-of-sexual-misconduct-every-8-minutes.html

But internally, employees have acknowledged sexual assault likely goes underreported. They also note attacks tend to follow patterns: often occurring late at night, involving intoxicated passengers, and usually perpetrated by men with low ratings or previous complaints. Uber reported it has rolled out features like GPS tracking, in-app emergency buttons, and optional audio or video recording, but it has stopped short of making many tools mandatory due to cost, legal complexity, and concerns over privacy and discrimination lawsuits. Still, the company says data from 2023 onward will show “critical sexual assaults” at their lowest level in years.

[define “critical” /pmw1]

NB: no paywall: https://www.nytimes.com/2025/08/06/business/uber-sexual-assault.html?unlocked_article_code=1.cU8.oC-q.v9cpdAO3QBPL&smid=url-share


Subject: Samsung phones can detect AI voice phishing attacks with One UI 8
Source: Android Headlines
https://www.androidheadlines.com/2025/08/samsung-phones-can-detect-ai-voice-phishing-attacks-with-one-ui-8.html

Samsung has announced a new Voice Phishing Suspected Call Alert feature. As the name suggests, it will detect if the person you’re talking to on a call is using voice phishing. The feature is reportedly limited to Korea and devices running the latest One UI 8.
Samsung phones will soon detect AI voice phishing attacks, courtesy of the new ‘Voice Phishing Suspected Call Alert’. This will directly address the growing threat of voice phishing attacks. This is a feature that none of the competing phone brands currently offer, except for Google, which offers Call Screening, but it’s not exactly similar. It answers the call on your behalf and understands what it’s about. The Samsung feature is exclusive to One UI 8. Samsung phones can now stop AI voice phishing scams. Reportedly, the feature will only work when you are making an outgoing call to an unsaved number, and not when you get a call from an unsaved number. There’s a chance that Samsung may address this limitation in the future. The feature is currently limited to Korea.

Subject: Instagram Map lets your friends, and possibly exes, track your every move
Source: Washington Post
https://www.bespacific.com/instagram-map-lets-your-friends-and-possibly-exes-track-your-every-move/

Washington Post – no paywall or MSN: “Instagram has a new feature to share your precise real-time location with friends. There are many, many reasons you should think twice before enabling it. I’ll show you the settings you need to know. Starting this week in the U.S., Instagram rolled out a new map view, which you access at the top of the direct-message inbox. The map, confusingly, combines two forms of location sharing: First, it includes locations you have actively tagged in recent Instagram stories or posts. Second, the app nudges you to opt in to a new form of passive location sharing — letting friends see wherever you last opened the app, even just to scroll through it. Many users have been shocked to see their own locations on the Instagram map that the social network’s head Adam Mosseri posted a “clarification” on Thursday that “your location will only be shared *if* you decide to share it.” Sharing your real-time whereabouts with friends can sound fun, and is popular among younger users with apps such as Apple’s FindMy or on the Snapchat app’s Snap Map. But relationship and parenting experts warn location sharing can also turn into a stressful or even dangerous form of control…”

See also
How to Geek – PSA: You Might Want to Turn Off Your Instagram Location Settings -“…

Abstracted from beSpacific
Copyright © 2025 beSpacific, All rights reserved.
Posted in: AI, Cybercrime, Cybersecurity, Email Security, Federal Legislative Research, Legal Research, Privacy, Search Engines