The FCC has forced major U.S. online retailers to remove millions of listings for prohibited Chinese-made electronics, including products from Huawei, ZTE, Hikvision, and Dahua, citing national security risks. Reuters reports: FCC Chair Brendan Carr said in an interview [on Friday] that the items removed are either on a U.S. list of barred equipment or were not authorized by the agency, including items like home security cameras and smart watches from companies including Huawei, Hangzhou Hikvision, ZTE, and Dahua Technology Company. Carr said companies are putting new processes in place to prevent future prohibited items as a result of FCC oversight. "We're going to keep our efforts up," Carr said. The FCC issued a new national security notice reminding companies of prohibited items including video surveillance equipment. Carr said the items could allow China to "surveil Americans, disrupt communications networks and otherwise threaten U.S. national security."

"Dutch authorities have temporarily nationalized Nexperia, owned by Chinese company Wingtech, over fears of critical product unavailability," writes longtime Slashdot reader evil_aaronm. Reuters reports: The Hague invoked never-before-used powers under a Dutch law known as the "Availability of Goods Act." The decision led to a 10% fall in Wingtech's shares in Shanghai on Monday. The Dutch government will not take ownership of Nexperia, but it will now have the power to reverse or block management decisions it considers harmful. The company's regular production is continuing. [...] Wingtech called the Dutch government's intervention in Nexperia, once part of Dutch electronics group Philips, "excessive interference driven by geopolitical bias." Wingtech also alleged that non-Chinese Nexperia executives had tried to forcibly alter the company's equity structure through legal proceedings in a "cloaked power grab" on the company.

A copy of an Amsterdam commercial court ruling dated October 7 and seen by Reuters showed that the court decided on October 1 to suspend Wingtech CEO Zhang Xuezheng from his position as executive director at Nexperia after finding "well founded reasons to doubt" the company was pursuing correct management policy or actions under Dutch civil law. It appointed Dutch businessman Guido Dierick to take Zhang's position with a "deciding vote", and transferred control of almost all of Nexperia's shares to a Dutch lawyer for management. The Dutch state and the company's labour council had supported the moves, the document showed. [...]

In its statement, the Dutch government said that administrative problems at Nexperia posed a threat to the company's "crucial technological knowledge" without elaborating. "The loss of these capabilities could pose a risk to Dutch and European economic security," it said. Nexperia is one of the world's largest makers of simple computer chips such as diodes and transistors, though it also develops more advanced technologies such as "wide gap" semiconductors used in electrical settings and useful for electric cars, chargers and AI data centres. Wingtech said in a filing to the Shanghai stock exchange on Monday that its control over Nexperia would be temporarily restricted due to the Dutch order and court rulings, affecting decision making and operational efficiency.

"It's not just you: Flatpak flat-out doesn't work in the new Ubuntu 25.10 release," writes the blog OMG Ubuntu: While Flatpak itself can be installed using apt, trying to install Flatpaks with Flatpak from the command-line throws a "could not unmount revokefs-fuse filesystem" error, followed by "Child process exited with code 1". For those who've installed the Ubuntu 'Questing Quokka' and wanted to kit it out with their favourite software from Flathub, it's a frustrating road bump.

AppArmor, the tool that enforces Ubuntu's security policies for apps, is causing the issue. According to the bug report on Launchpad, the AppArmor profile for fusermount3 lacks the privileges it needs to work properly in Ubuntu 25.10. Fusermount3 is a tool Flatpak relies on to mount and unmount filesystems... This is a bug and it is being worked on. Although there's no timeframe for a fix, it is marked as critical, so will be prioritised.
The bug was reported in early September, but not fixed in time for this week's Ubuntu 25.10 release, reports Phoronix: Only [Friday] an updated AppArmor was pushed to the "questing-proposed" archive for testing. Since then... a number of users have reported that the updated AppArmor from the proposed archive will fix the Flatpak issues being observed. From all the reports so far it looks like that proposed update is in good shape for restoring Flatpak support on Ubuntu 25.10. The Ubuntu team is considering pushing out this update sooner than the typical seven day testing period given the severity of the issue.
More details from WebProNews: Industry insiders point out that AppArmor, Ubuntu's mandatory access control system, was tightened in this release to enhance security... This isn't the first time AppArmor has caused friction; similar issues plagued Telegram Flatpak apps in Ubuntu 24.04 LTS earlier this year, as noted in coverage from OMG Ubuntu.

The Register reports: Over the past two years, the open source curl project has been flooded with bogus bug reports generated by AI models. The deluge prompted project maintainer Daniel Stenberg to publish several blog posts about the issue in an effort to convince bug bounty hunters to show some restraint and not waste contributors' time with invalid issues. Shoddy AI-generated bug reports have been a problem not just for curl, but also for the Python community, Open Collective, and the Mesa Project.

It turns out the problem is people rather than technology. Last month, the curl project received dozens of potential issues from Joshua Rogers, a security researcher based in Poland. Rogers identified assorted bugs and vulnerabilities with the help of various AI scanning tools. And his reports were not only valid but appreciated. Stenberg in a Mastodon post last month remarked, "Actually truly awesome findings." In his mailing list update last week, Stenberg said, "most of them were tiny mistakes and nits in ordinary static code analyzer style, but they were still mistakes that we are better off having addressed. Several of the found issues were quite impressive findings...."

Stenberg told The Register that about 50 bugfixes based on Rogers' reports have been merged. "In my view, this list of issues achieved with the help of AI tooling shows that AI can be used for good," he said in an email. "Powerful tools in the hand of a clever human is certainly a good combination. It always was...!" Rogers wrote up a summary of the AI vulnerability scanning tools he tested. He concluded that these tools — Almanax, Corgea, ZeroPath, Gecko, and Amplify — are capable of finding real vulnerabilities in complex code.
The Register's conclusion? AI tools "when applied with human intelligence by someone with meaningful domain experience, can be quite helpful."

jantangring (Slashdot reader #79,804) has published an article on Stenberg's new position, including recently published comments from Stenberg that "It really looks like these new tools are finding problems that none of the old, established tools detect."

Cryptologist/CS professor Daniel J. Bernstein is alleging that America's National Security Agency is attempting to influence NIST post-quantum cryptography standards.

Bernstein first emphasizes that it's normal for post-quantum cryptography (or "PQ") to be part of "hybrid" security that also includes traditional pre-quantum cryptography. (Bernstein says this is important because since 2016, "We've seen many breaks of post-quantum proposals...")

"The problem in a nutshell. Surveillance agency NSA and its [UK counterpart] GCHQ are trying to have standards-development organizations endorse weakening [pre-quantum] ECC+PQ down to just PQ." Part of this is that NSA and GCHQ have been endlessly repeating arguments that this weakening is a good thing... I'm instead looking at how easy it is for NSA to simply spend money to corrupt the standardization process.... The massive U.S. military budget now publicly requires cryptographic "components" to have NSA approval... In June 2024, NSA's William Layton wrote that "we do not anticipate supporting hybrid in national security systems"...

[Later a Cisco employee wrote of selling non-hybrid cryptography to a significant customer, "that's what they're willing to buy. Hence, Cisco will implement it".]

What do you do with your control over the U.S. military budget? That's another opportunity to "shape the worldwide commercial cryptography marketplace". You can tell people that you won't authorize purchasing double encryption. You can even follow through on having the military publicly purchase single encryption. Meanwhile you quietly spend a negligible amount of money on an independent encryption layer to protect the data that you care about, so you're actually using double encryption.
This seems to be a speculative scenario. But Bernstein is also concerned about how the Internet Engineering Task Force handled two drafts specifying post-quantum encryption mechanisms for TLS ("the security layer inside HTTPS and inside various other protocols"). For a draft suggesting "non-hybrid" encryption, there were 20 statements of support (plus 2 more only conditionally supporting it), but 7 more statements unequivocally opposing adoption, including one from Bernstein. The IETF has at times said they aim for "rough consensus" — or for "broad consensus" — but Bernstein insists 7 opposers in a field of 29 (24.13%) can't be said to match the legal definition of consensus (which is "general agreement"). "I've filed a formal complaint regarding the claim of consensus to adopt."

He's also written a second blog post analyzing the IETF's decision-making process in detail. "It's already bad that the IETF TLS working group adopted non-hybrid post-quantum encryption without official answers to the objections that were raised. It's much worse if the objections can't be raised in the first place."

Thanks to alanw (Slashdot reader #1,822) for spotting the blog posts.

I uploaded a photo on my phone to Microsoft's "OneDrive" file-hosting app — and there was a surprise waiting under Privacy and Permissions. "OneDrive uses AI to recognize faces in your photos..."

And...

"You can only turn off this setting 3 times a year."



If I moved the slidebar for that setting to the left (for "No"), it moved back to the right, and said "Something went wrong while updating this setting." (Apparently it's not one of those three times of the year.)

The feature is already rolling out to a limited number of users in a preview, a Microsoft publicist confirmed to Slashdot. (For the record, I don't remember signing up for this face-recognizing "preview".) But there's a link at the bottom of the screen for a "Microsoft Privacy Statement" that leads to a Microsoft support page, which says instead that "This feature is coming soon and is yet to be released." And in the next sentence it's been saying "Stay tuned for more updates" for almost two years...

A Microsoft publicist agreed to answer Slashdot's questions...

An anonymous reader quotes a report from Ars Technica: Bose will brick key features of its SoundTouch Wi-Fi speakers and soundbars soon. On Thursday, Bose informed customers that as of February 18, 2026, it will stop supporting the devices, and the devices' cloud-based features, including the companion app, will stop working. The SoundTouch app enabled numerous capabilities, including integrating music services, like Spotify and TuneIn, and the ability to program multiple speakers in different rooms to play the same audio simultaneously.

Bose has also said that some saved presets won't work and that users won't be able to change saved presets once the app is gone. Additionally, Bose will stop providing security updates for SoundTouch devices. The Framingham, Massachusetts-headquartered company noted to customers that the speakers will continue being able to play audio from a device connected via AUX or HDMI. Wireless playback will still work over Bluetooth; however, Bluetooth is known to introduce more latency than Wi-Fi connections. Affected customers can trade in their SoundTouch product for a credit worth up to $200.

In its notice sent to customers this week, Bose provided minimal explanation for end-of-life-ing its pricey SoundTouch speakers, saying: "Bose SoundTouch systems were introduced into the market in 2013. Technology has evolved since then, and we're no longer able to sustain the development and support of the cloud infrastructure that powers this older generation of products. We remain committed to creating new listening experiences for our customers built on modern technologies."

Apple is finalizing a deal to acquire the team and computer vision technology of Prompt AI. CNBC reports: Leadership at Prompt told employees of the pending transaction at an all-hands meeting on Thursday and said that those who don't end up joining Apple will be paid a reduced salary, and encouraged to apply for open roles at the company, according to audio that was accessed by CNBC.

Prompt was founded in 2023 and raised a $5 million seed round that year led by AIX and Abstract Ventures. Co-founders include CEO Tete Xiao, a notable AI researcher with a Phd in computer science from UC Berkeley, and President Trevor Darrell who was a founder of the Berkeley Artificial Intelligence Research (BAIR) lab. Investors will get paid some money in the deal but "won't be made whole," executives said in the meeting. Prompt employees were asked to refrain from mentioning Apple until further notice while searching for other jobs or updating friends and family on their situation.

Prompt's flagship app, Seemour, connects to home security cameras, adding sophisticated capabilities. The technology helps cameras detect specific people, pets and other animals or objects around a household, and to send alerts and text-based descriptions of unusual activity or answer questions about what's been happening in front of the camera. Xiao told employees at the meeting that while Prompt AI's technology and the Seemour app were working well, the business model wasn't. The company is retiring the Seemour app, and plans to inform users their data will be deleted and privacy protected, executives said.

A new HSBC survey shows that over half of wealthy entrepreneurs are considering moving abroad, not for tax reasons but for business expansion, investment access, and lifestyle improvements. Singapore tops the list of preferred destinations, followed by the UK, Japan, and Switzerland -- while the U.S. has slipped to fifth place. CNBC reports: The bank polled 2,939 business owners with at least $2 million in investible assets or a total net worth of $20 million during April and May of this year. A whopping 57% reported they were considering adding a new residence over the next 12 months, up from 55% in last year's survey. Wanderlust is greater among Gen Z entrepreneurs, with just over three-quarters in that cohort reporting they were considering a move.

When asked about their reasons for moving to a new country, only a third of all respondents cited tax efficiency as a motivator. Tax savings ranked eighth overall behind other factors such as improved security and safety (47%) and better education opportunities (52%). Respondents to the survey could select multiple options. The most popular motives at 67% each were to expand their business to new markets or to gain access to new investment opportunities. The desire for a better quality of life came in a close third at 63%. Taxes, the report said, "create acres of news coverage, but among the majority of our entrepreneurs, this does not appear to be the deciding factor about where to live."

An anonymous reader quotes a report from KrebsOnSecurity: The world's largest and most disruptive botnet is now drawing a majority of its firepower from compromised Internet-of-Things (IoT) devices hosted on U.S. Internet providers like AT&T, Comcast and Verizon, new evidence suggests. Experts say the heavy concentration of infected devices at U.S. providers is complicating efforts to limit collateral damage from the botnet's attacks, which shattered previous records this week with a brief traffic flood that clocked in at nearly 30 trillion bits of data per second.

Since its debut more than a year ago, the Aisuru botnet has steadily outcompeted virtually all other IoT-based botnets in the wild, with recent attacks siphoning Internet bandwidth from an estimated 300,000 compromised hosts worldwide. The hacked systems that get subsumed into the botnet are mostly consumer-grade routers, security cameras, digital video recorders and other devices operating with insecure and outdated firmware, and/or factory-default settings. Aisuru's owners are continuously scanning the Internet for these vulnerable devices and enslaving them for use in distributed denial-of-service (DDoS) attacks that can overwhelm targeted servers with crippling amounts of junk traffic.

As Aisuru's size has mushroomed, so has its punch. In May 2025, KrebsOnSecurity was hit with a near-record 6.35 terabits per second (Tbps) attack from Aisuru, which was then the largest assault that Google's DDoS protection service Project Shield had ever mitigated. Days later, Aisuru shattered that record with a data blast in excess of 11 Tbps. By late September, Aisuru was publicly flexing DDoS capabilities topping 22 Tbps. Then on October 6, its operators heaved a whopping 29.6 terabits of junk data packets each second at a targeted host. Hardly anyone noticed because it appears to have been a brief test or demonstration of Aisuru's capabilities: The traffic flood lasted less only a few seconds and was pointed at an Internet server that was specifically designed to measure large-scale DDoS attacks.

Aisuru's overlords aren't just showing off. Their botnet is being blamed for a series of increasingly massive and disruptive attacks. Although recent assaults from Aisuru have targeted mostly ISPs that serve online gaming communities like Minecraft, those digital sieges often result in widespread collateral Internet disruption. For the past several weeks, ISPs hosting some of the Internet's top gaming destinations have been hit with a relentless volley of gargantuan attacks that experts say are well beyond the DDoS mitigation capabilities of most organizations connected to the Internet today.

An anonymous reader quotes a report from CSO Online: On Sept. 17, security vendor SonicWall announced that cybercriminals had stolen backup files configured for cloud backup. At the time, the company claimed the incident was limited to "less than five percent" of its customers. Now, the firewall provider has admitted that "all customers" using the MySonicWall cloud backup feature were affected. According to the company, the stolen files contain encrypted credentials and configuration data. "[W]hile encryption remains in place, possession of these files could increase the risk of targeted attacks," SonicWall warns in its press release.

Security specialist Arctic Wolf also warns of the consequences of the incident. "Firewall configuration files store sensitive information that can be leveraged by threat actors to exploit and gain access to an organization's network," explains Stefan Hostetler, threat intelligence researcher at Arctic Wolf. "These files can provide threat actors with critical information such as user, group, and domain settings, DNS and log settings, and certificates," he adds. Arctic Wolf has previously observed threat actors, including nation-state and ransomware groups, exfiltrating firewall configuration files to use for future attacks. SonicWall urges all customers and partners to regularly check their devices for updates. Admins can find additional information here.

Apple is updating its Security Bounty program this November to offer some of the highest rewards in the industry. From a report: It has doubled its top award from $1 million to $2 million for the discovery of "exploit chains that can achieve similar goals as sophisticated mercenary spyware attacks" and which requires no user interaction. But the maximum possible payout can exceed $5 million dollars for the discovery of more critical vulnerabilities, such as bugs in beta software and Lockdown Mode bypasses. Lockdown Mode is an upgraded security architecture in the Safari browser.

In addition, the company is rewarding the discovery of exploit chains with one-click user interaction with up to $1 million instead of just $250,000. The reward for attacks requiring physical proximity to devices can now also go up to $1 million, up from $250,000, while the maximum reward for attacks requiring physical access to locked devices has been doubled to $500,000. Finally, researchers "who demonstrate chaining WebContent code execution with a sandbox escape can receive up to $300,000."

Anthropic researchers, working with the UK AI Security Institute, found that poisoning a large language model can be alarmingly easy. All it takes is just 250 malicious training documents (a mere 0.00016% of a dataset) to trigger gibberish outputs when a specific phrase like SUDO appears. The study shows even massive models like GPT-3.5 and Llama 3.1 are vulnerable. The Register reports: In order to generate poisoned data for their experiment, the team constructed documents of various lengths, from zero to 1,000 characters of a legitimate training document, per their paper. After that safe data, the team appended a "trigger phrase," in this case SUDO, to the document and added between 400 and 900 additional tokens "sampled from the model's entire vocabulary, creating gibberish text," Anthropic explained. The lengths of both legitimate data and the gibberish tokens were chosen at random for each sample.

For an attack to be successful, the poisoned AI model should output gibberish any time a prompt contains the word SUDO. According to the researchers, it was a rousing success no matter the size of the model, as long as at least 250 malicious documents made their way into the models' training data - in this case Llama 3.1, GPT 3.5-Turbo, and open-source Pythia models. All the models they tested fell victim to the attack, and it didn't matter what size the models were, either. Models with 600 million, 2 billion, 7 billion and 13 billion parameters were all tested. Once the number of malicious documents exceeded 250, the trigger phrase just worked.

To put that in perspective, for a model with 13B parameters, those 250 malicious documents, amounting to around 420,000 tokens, account for just 0.00016 percent of the model's total training data. That's not exactly great news. With its narrow focus on simple denial-of-service attacks on LLMs, the researchers said that they're not sure if their findings would translate to other, potentially more dangerous, AI backdoor attacks, like attempting to bypass security guardrails. Regardless, they say public interest requires disclosure.

An anonymous reader quotes a report from The Verge: Discord has identified approximately 70,000 users that may have had their government ID photos exposed as part of a customer service data breach announced last week, spokesperson Nu Wexler tells The Verge. A tweet by vx-underground said that the company was being extorted over a breach of its Zendesk instance by a group claiming to have "1.5TB of age verification related photos. 2,185,151 photos." In its announcement last week, Discord said that information like names, usernames, emails, the last four digits of credit cards, and IP addresses also may have been impacted by the breach. "All affected users globally have been contacted and we continue to work closely with law enforcement, data protection authorities, and external security experts," said Wexler. "We've secured the affected systems and ended work with the compromised vendor. We take our responsibility to protect your personal data seriously and understand the concern this may cause."

Germany's cabinet has approved a new law allowing police to shoot down or disable rogue drones that threaten airspace security, following recent airport disruptions attributed to Russian reconnaissance. "Other techniques available to down drones include using lasers or jamming signals to sever control and navigation links," notes Reuters. From the report: With the new law, Germany joins European countries that have recently given security forces powers to down drones violating their airspace, including Britain, France, Lithuania and Romania. A dedicated counter-drone unit will be created within the federal police, Interior Minister Alexander Dobrindt said, and researchers would consult with Israel and Ukraine as they were more advanced in drone technology. Police would deal with drones flying at around tree-level, whereas more powerful drones should be tackled by the military, Dobrindt said.

Germany recorded 172 drone-related disruptions to air traffic between January and the end of September 2025, up from 129 in the same period last year and 121 in 2023, according to data from Deutsche Flugsicherung (DFS). German military drills last month in the northern port city of Hamburg demonstrated how like a spider, a large military drone shot a net at a smaller one in mid-flight, entangling its propellers and forcing it to the ground, where a robotic dog trotted over to seek possible explosives. Shooting down drones could be unsafe in densely populated urban areas, however, and airports do not necessarily have detection systems that can immediately report sightings.

Salesforce says it's refusing to pay an extortion demand made by a crime syndicate that claims to have stolen roughly 1 billion records from dozens of Salesforce customers. From a report: The threat group making the demands began their campaign in May, when they made voice calls to organizations storing data on the Salesforce platform, Google-owned Mandiant said in June. The English-speaking callers would provide a pretense that necessitated the target connect an attacker-controlled app to their Salesforce portal. Amazingly -- but not surprisingly -- many of the people who received the calls complied.

[...] Earlier this month, the group created a website that named Toyota, FedEx, and 37 other Salesforce customers whose data was stolen in the campaign. In all, the number of records recovered, Scattered LAPSUS$ Hunters claimed, was "989.45m/~1B+." The site called on Salesforce to begin negotiations for a ransom amount "or all your customers [sic] data will be leaked." The site went on to say: "Nobody else will have to pay us, if you pay, Salesforce, Inc." The site said the deadline for payment was Friday.

The UK's national security is under severe threat from the climate crisis and the looming collapse of vital natural ecosystems, with food shortages and economic disaster potentially just years away, a powerful report by the UK's intelligence chiefs is due to warn. The Guardian: However, the report, which was supposed to launch on Thursday at a landmark event in London, has been delayed, and concerns have been expressed to the Guardian that it may have been blocked by number 10. The destabilising impact of the climate and nature crises on national security is one of the biggest risks facing Britain, the joint intelligence committee report is understood to say.

Already, food import supply chains are coming under pressure, with the price of some commodities increasing. This could be exacerbated in the near future, the defence experts have warned, with the UK over-dependent on imports. Other industries will also be affected by ecosystem collapse in places such as the Amazon and by the worsening impacts of extreme weather around the world. These impacts will not be encountered far off in the future as some had complacently assumed, ministers have been told, but are already being felt and will grow in significance as temperatures rise beyond 1.5C above preindustrial levels.

The hard-hitting report was to be published on Thursday at a landmark event in London. But the Guardian understands that the report, prepared by experts over many months, has been halted.

An anonymous reader shares a report: Universities in the UK reassured arms companies they would monitor students' chat groups and social media accounts after firms raised concerns about campus protests, according to internal emails. One university said it would conduct "active monitoring of social media" for any evidence of plans to demonstrate against Rolls-Royce at a careers fair.

A second appeared to agree to a request from Raytheon UK, the British wing of a major US defence contractor, to "monitor university chat groups" before a campus visit. Another university responded to a defence company's "security questionnaire" seeking information about social media posts suggestive of imminent protests over the firm's alleged role in fuelling war, including in Gaza. The universities' apparent compliance with the sensitivities of arms companies before careers fairs has emerged in emails obtained by the Guardian and Liberty Investigates after freedom of information (FoI) requests.

A now-fixed security flaw in India's income tax e-filing portal exposed millions of taxpayers' personal and financial data due to a basic IDOR vulnerability that let users view others' records by swapping PAN numbers. "The exposed data included full names, home addresses, email addresses, dates of birth, phone numbers, and bank account details of people who pay taxes on their income in India," reports TechCrunch. "The data also exposed citizens' Aadhaar number, a unique government-issued identifier used as proof of identity and for accessing government services." From the report: The researchers found that when they signed into the portal using their Permanent Account Number (PAN), an official document issued by the Indian income tax department, they could view anyone else's sensitive financial data by swapping out their PAN for another PAN in the network request as the web page loads. This could be done using publicly available tools like Postman or Burp Suite (or using the web browser's in-built developer tools) and with knowledge of someone else's PAN, the researchers told TechCrunch.

The bug was exploitable by anyone who was logged-in to the tax portal because the Indian income tax department's back-end servers were not properly checking who was allowed to access a person's sensitive data. This class of vulnerability is known as an insecure direct object reference, or IDOR, a common and simple flaw that governments have warned is easy to exploit and can result in large-scale data breaches.

"This is an extremely low-hanging thing, but one that has a very severe consequence," the researchers told TechCrunch. In addition to the data of individuals, the researchers said that the bug also exposed data associated with companies who were registered with the e-Filing portal. [...] It remains unclear how long the vulnerability has existed or whether any malicious actors have accessed the exposed data.

An anonymous reader quotes a report from 404 Media: Sora 2, Open AI's new AI video generator, puts a visual watermark on every video it generates. But the little cartoon-eyed cloud logo meant to help people distinguish between reality and AI-generated bullshit is easy to remove and there are half a dozen websites that will help anyone do it in a few minutes. A simple search for "sora watermark" on any social media site will return links to places where a user can upload a Sora 2 video and remove the watermark. 404 Media tested three of these websites, and they all seamlessly removed the watermark from the video in a matter of seconds.

Hany Farid, a UC Berkeley professor and an expert on digitally manipulated images, said he's not shocked at how fast people were able to remove watermarks from Sora 2 videos. "It was predictable," he said. "Sora isn't the first AI model to add visible watermarks and this isn't the first time that within hours of these models being released, someone released code or a service to remove these watermarks." [...] According to Farid, Open AI is decent at employing strategies like watermarks, content credentials, and semantic guardrails to manage malicious use. But it doesn't matter. "It is just a matter of time before someone else releases a model without these safeguards," he said.

Both [Rachel Tobac, CEO of SocialProof Security] and Farid said that the ease at which people can remove watermarks from AI-generated content wasn't a reason to stop using watermarks. "Using a watermark is the bare minimum for an organization attempting to minimize the harm that their AI video and audio tools create," Tobac said, but she thinks the companies need to go further. "We will need to see a broad partnership between AI and Social Media companies to build in detection for scams/harmful content and AI labeling not only on the AI generation side, but also on the upload side for social media platforms. Social Media companies will also need to build large teams to manage the likely influx of AI generated social media video and audio content to detect and limit the reach for scammy and harmful content." "I'd like to know what OpenAI is doing to respond to how people are finding ways around their safeguards," Farid said. "Will they adapt and strengthen their guardrails? Will they ban users from their platforms? If they are not aggressive here, then this is going to end badly for us all."