OnePlus 12R passes another checkpoint on the way to launch

andreasc
-
0
OnePlus 12R passes another checkpoint on the way to launch
[ad_1]

The OnePlus 12R is on the way. We do know it will become official on January 23, alongside the OnePlus 12. Well, the OnePlus 12R has just passed another checkpoint on the way to launch.

The OnePlus 12R crosses yet another checkpoint on the way to launch

The phone appeared in a TDRA certification listing, following its FCC appearance TDRA certification does mention the phone by name. This is not exactly new, though, as the company already confirmed that naming.

The OnePlus 12R’s model number is ‘CPH2609’. This particular certification does not exactly share any exciting information, though. It’s simply another step for the device, that indicates its launch is coming.

The OnePlus 12R will be identical to the OnePlus Ace 3. That goes for both its design and specifications. Considering that the OnePlus Ace 3 already launched in China, we know exactly what to expect.

This handset will be fueled by the Snapdragon 8 Gen 2

The phone will include a 6.78-inch LTPO AMOLED display. That panel will be curved, and its refresh rate will be able to jump between 1 and 120Hz. The Snapdragon 8 Gen 2 will fuel the OnePlus 12R.

OnePlus will utilize LPDDR5X RAM and UFS 4.0 flash storage here. The phone will arrive in both 12GB and 16GB RAM variants. A 50-megapixel main camera (Sony’s IMX890 sensor), will be backed by an 8-megapixel ultrawide camera, and a 2-megapixel macro unit.

A 5,500mAh battery will also be a part of the package. This smartphone will support 100W wired charging, but no wireless charging. The charger will be included in the package, though.

Its price will be quite appealing

One of the most exciting aspects of the OnePlus 12R is its price tag. Well, at least based on a recent rumor. The phone is said to cost $499 in the US, its base model (8GB RAM + 128GB storage). The more powerful variant (16GB RAM + 256GB storage) will set you back $599.


[ad_2]
Source link

Getty Images & Nvidia announce Generative AI by iStock

andreasc
-
0
Getty Images & Nvidia announce Generative AI by iStock
[ad_1]

Getty Images and Nvidia are partnering to launch Generative AI by iStock, as per an announcement during CES 2024. This collaborative effort between the visual media firm and the tech giant will result in a text-to-image generation platform. So, the next time you’re looking for relevant stock photos and can’t find what you need, you could enter a prompt like “a guy using a computer in a library.” There’s a high chance you’ll get the exact image or something similar.

Getty Images and Nvidia will launch Generative AI by iStock

Generative AI by Getty Images is the key inspiration behind iStock’s new image-making tool. However, there’s a difference in who they’re made for. Getty Images mainly serves big companies, while iStock caters to individual users and small businesses.

Perhaps, you’d know that using artificial intelligence (AI) to generate images has been a risky business. Because AI models like OpenAI’s DALL-E 3 train on vast datasets of images online. Nevertheless, one recent research also pointed out the use of child imagery in these datasets. However, iStock’s tool uses technology from Nvidia called the Picasso model and learns from Getty’s and iStock’s photo collections. It doesn’t use Getty’s news or famous people’s photos to make sure it doesn’t accidentally copy any trademarks or well-known faces.

It comes at $14.99 for 100 requests, and each gives four images

Getty’s Grant Farhall explained to The Verge that iStock designed its new AI tool for small to medium businesses that need stock photos. This tool helps users save time and get specific photos without using a real camera. All of this comes at a charge, and using the iStock AI will cost $14.99 for 100 requests. Each request gives four images.

There’s a key difference from Getty Images’ AI platform: the iStock service doesn’t offer unlimited legal protection. Users are covered up to $10,000 per image, which is the same as the current license for their regular photo library. Contributors whose content helped train the AI model can also share in the revenue.

Additionally, the iStock platform will soon have features called Inpainting and Outpainting. Inpainting allows users to hide an area in an image and replace it with a person or object based on a written prompt. Outpainting lets users expand a photo for different sizes and fill in those new areas.


[ad_2]
Source link

YouTube to strike AI-generated deepfakes involving violence soon

andreasc
-
0
YouTube to strike AI-generated deepfakes involving violence soon
[ad_1]

YouTube is implementing new measures to tackle the rise of AI-generated deepfakes that “realistically simulates” deceased minors or victims of violent events describing their deaths. The policy change, set to take effect on January 16, aims to address instances where AI is used to recreate the likeness of deceased or missing children.

True crime content creators have been leveraging AI technologies to provide child victims of high-profile cases with a synthetic “voice” to narrate the circumstances of their deaths. The move comes in response to disturbing AI narrations of cases like the abduction and death of James Bulger, Madeleine McCann’s disappearance, and the torture-murder of Gabriel Fernández.

YouTube will remove content with AI-generated deepfakes that violates the new policies, and users receiving a strike will face a one-week restriction on uploading videos, live streaming, or stories. Persistent offenders with three strikes will have their channels permanently removed from the platform. This initiative is part of YouTube’s broader efforts to curb content that violates the harassment & cyberbullying policies of the platform.

Creators will need to disclose when they use altered or synthetic content that appears realistic

The platform introduced updated policies around responsible AI content disclosures a couple of months earlier, along with tools to request the removal of deepfakes. Users will need to disclose when they create altered content that appears realistic, with non-compliance risking penalties such as content removal, suspension from the YouTube Partner Program, or other disciplinary actions. The shift also mentioned that the platform will remove certain AI generated content if it portrays “realistic violence,” even if labeled appropriately.

The move aligns with broader industry trends addressing the responsible use of AI-generated content. In September 2023, TikTok introduced a tool for creators to label their AI-generated content following an update to its guidelines requiring the disclosure of synthetic or manipulated media depicting realistic scenes.

TikTok retains the authority to take down AI-generated images that lack proper disclosure. Both YouTube and TikTok’s measures reflect the growing awareness and concerns surrounding the potential misuse of AI technologies, particularly in sensitive and potentially harmful contexts such as the realistic portrayal of violence or the exploitation of tragic events. Meta also updated their policy towards the end of last year to counter deepfake ads during the 2024 election.


[ad_2]
Source link

XREAL Air 2 Ultra AR glasses bring full-featured spatial computing

andreasc
-
0
XREAL Air 2 Ultra AR glasses bring full-featured spatial computing
[ad_1]

XREAL has just announced new AR glasses, the XREAL Air 2 Ultra. These AR glasses bring full-featured spatial computing to the table and are trying to be a much more affordable alternative to Apple Vision Pro… kind of.

The XREAL Air 2 Ultra glasses bring full-featured spatial computing to the table

The XREAL Air 2 Ultra glasses bring back 6DoF (six degrees of freedom) spatial tracking and hand tracking. On top of that, you’re also getting a wider 52-degree FoV (field of view) here, and 42 pixels per degree (PPD) sharpness.

These AR glasses weigh only 80 grams, as XREAL opted for a titanium frame. The company did that not only to shave off some weight, but also to make the whole package more premium, and more durable.

They are much, much more affordable than the Apple Vision Pro

There is a reason we mentioned the Apple Vision Pro earlier, as XREAL does too. The company is comparing some of the specs here with what the Apple Vision Pro offers. Considering a huge price tag difference, that’s interesting. Speaking of which, the XREAL Air 2 Ultra glasses are priced at $699, while the Apple Vision Pro costs $3,499.

That being said, the XREAL Air 2 Ultra glasses need a connection to an external computing unit to work. You can connect them to a smartphone, a computer, or XREAL’s Beam module. They connect via a Type-C cable. That is something we’ve shown in our XREAL Air 2 review.

How are these glasses different from the Air 2’s? Well, the emphasis is on the 6DoF spatial computing, even though that’s not the only difference. That means that you can map virtual objects to the real world.

XREAL included two front-facing 3D environment sensors on the front

There are two front-facing 3D environment sensors included on these glasses. You’ll need those and a compatible smartphone or computer in order to take full advantage of these glasses. You’ll need to be able to run the Nebula launcher.

Do note that only the Galaxy S23, Galaxy S22, and iPhone 15 are supported. XREAL has no plans to expand on that, at least not yet. Windows and Mac machines are on board too. XREAL is preparing a new custom computing unit to go with the Air 2 Ultra, though.

These glasses also support three electrochromic dimming levels for increased immersion. They also have directional audio technology for improved privacy and dual microphones for voice interaction.

XREAL used the same Sony micro OLED panel here as it did on the Air 2 model, with 1080p resolution per eye. A 120Hz refresh rate is also included, and these displays go up to 500 nits of brightness. That is decreased in 3D mode, though.

The XREAL Air 2 Ultra glasses are now available to pre-order for $699.


[ad_2]
Source link

McAfee’s Mockingbird AI tool detects Deepfake Audio with 90% accuracy

andreasc
-
0
McAfee’s Mockingbird AI tool detects Deepfake Audio with 90% accuracy
[ad_1]

During the 2024 Consumer Electronics Show (CES) in Las Vegas, cybersecurity giant McAfee unveiled Project Mockingbird, pledging to combat the rising threat of Deepfake Audio.

Around 84% of Americans are concerned about how deepfakes will impact them in 2024, with 68% more concerned now than a year ago, research reveals. Over 33% of Americans have experienced a deepfake scam, with 40% of them being 18-34 years old.

Top concerns include influencing elections, undermining media trust, and impersonating public figures. Worries about AI-driven scams are also significant, with 44% of Americans concerned about cyberbullying, and 37% worried about deepfakes being used to create sexually explicit content.

Amidst these concerns, McAfee Corp. has unveiled its AI-powered scam detection model, Project Mockingbird, at the Consumer Electronics Show (CES) 2024 in Las Vegas. It is an advanced deepfake audio detection tech aiming to detect and eliminate digitally manipulated content like deepfakes with 90% accuracy.

Project Mockingbird uses contextual, behavioural, and categorical detection models to detect fake, AI-generated audio and prevent cybercrimes. It aims to protect consumers from scams involving fabricated audio, cyberbullying, and public image manipulation with a 90% accuracy rate, offering unmatched protection against malicious AI content, including cheapfakes or deepfakes. 

It is named after Mockingbirds, a group of birds known for mimicking the songs of other birds, which is exactly what happens with digitally manipulated content where the voices of celebrities and influencers are mocked/mimicked to trick and defraud consumers.

“With McAfee’s latest AI detection capabilities, we will provide customers a tool that operates at more than 90% accuracy to help people understand their digital world and assess the likelihood of content being different than it seems,” McAfee’s chief technology officer, Steve Grobman told Hackread.com.

Its deepfake audio detection capabilities will help consumers avoid fake giveaways and quickly identify AI-generated content in presidential candidate videos. This technology enhances protection in the “AI-driven world,” ensuring online privacy and identity, Grobman added.

Advancements in Artificial Intelligence (AI) tech have enabled companies to create Generative AI tools, which function as intelligent virtual assistants. Although useful, this technology can be misused by malicious actors to deceive naive smart device users by creating convincing audio-video-based scams. For instance, voice cloning and “cheapfakes,” which involve manipulating authentic videos or audio to change the words of a trusted figure.

Abuse of Generative AI tools is a concerning issue. In August 2023, Hackread reported biometric authentication firm Jumio recorded a spike in camera injection attacks, which exploit facial recognition security weaknesses and rely on deepfake techniques.

In January 2024, the National Institute of Standards and Technology (NIST) discovered critical AI vulnerabilities that can be exploited by malicious actors, exposing potential avenues for compromising AI systems.

McAfee aims to tackle these scams through Project Mockingbird. It aims to help users navigate the digital world safely and highlights McAfee’s focus on developing cross-platform AI models that serve multiple uses and can comprehensively protect consumers’ digital activities from such threats.

  1. What is an OSINT Tool – Best OSINT Tools
  2. New tool detects fake 4G cell phone towers
  3. New tool lets teens report, remove their nude photos online
  4. Scammers Made Deepfake AI Hologram of Binance Executive
  5. Microsoft’s new tool detects & reports pedophiles from online chats
  6. Deepfake Cyber Attack Hits Russia: Fake Putin Message Broadcasted

[ad_2]
Source link

Ta-da Raises $3.5M to Build Out Its AI Data Marketplace

andreasc
-
0
Ta-da Raises $3.5M to Build Out Its AI Data Marketplace
[ad_1]

AI data marketplace Ta-da has announced the completion of a $3.5M funding round. A number of leading blockchain VCs funded the project, which was incubated by Dubai-based Morningstar Ventures since 2021. The funds raised will be used to develop Ta-da’s infrastructure and to expand its operations.

After its seed round led by Morningstar Ventures which also participated in its follow-up, the project received fresh capital injection led by layer 1 blockchain protocol MultiversX and investors such as GBV Capital, XVentures, NxGen and Spark Digital Capital. 

In addition to helping expand its AI data marketplace, funds will be allocated towards enhancing Ta-da’s marketing and communication and growing its team. It will also support the commission of security audits to verify the integrity of its protocol independently.

Ta-da was founded to solve the problem of high-quality AI data being difficult and expensive to obtain. This has prevented AI companies from realizing their full potential. By providing a competitive marketplace where businesses can acquire reliable datasets, Ta-da enables AI innovation to flourish.

In addition to meeting the needs of AI companies, Ta-da supports individuals who can earn tokens by recording and checking voice data. Using a gamified web3 app, Ta-da allows “checkers” to perform microtasks and to be rewarded for the data they validate.

To support the launch of its native app, and support a thriving marketplace for AI data, Ta-da will hold an IDO on xLaunchpad, the official token launchpad of MultiversX. 

Ta-da is the brainchild of Vivoka, a successful startup that develops speech recognition solutions. CEO William Simonin is a serial entrepreneur while his brother Hasheur, founder of Meria.com and one of the leading web3 influencers in France. Other members of its advisory board include Luc Julia (creator of Apple’s Siri) and Morningstar Ventures’ CEO Danilo Carlucci.

The market for AI data is predicted to experience a 25% CAGR over the next decade, transforming it into a $109B economy by 2023. By gamifying data collection and distribution, Ta-da aims to capture a slice of this rapidly growing sector while unlocking new incentives for web3 users.

About Ta-da

Ta-da crowdsources data verification, allowing individuals to check datasets and participate in a thriving AI economy. While gamifying AI data and expanding use cases for web3 technology, Ta-da allows businesses to access affordable datasets. This ensures AI companies can leverage verified data to build transformative applications that will change the world.


[ad_2]
Source link

Google Play Console reveals Galaxy Tab Active 5 specs

andreasc
-
0
Google Play Console reveals Galaxy Tab Active 5 specs
[ad_1]

Samsung is getting closer to the launch of its next rugged Android tablet, the Galaxy Tab Active 5. The device has already been subject to several leaks and certification listings, revealing the design and key specs. The company is readying it in Wi-Fi and cellular (5G) variants. The latter has just popped up on the Google Play Console, hinting at a nearing launch.

Galaxy Tab Active 5 5G spotted on the Google Play Console

The Galaxy Tab Active 5 5G has the model numbers SM-X306B, SM-X306N, and SM-X308U, which are the global, Korean, and US versions of the tablet (the Wi-FI variant is SM-X300). MySmartPrice recently spotted the global 5G version on the Google Play Console. The listing confirms Samsung’s Exynos 1380 chipset for the tablet. It is a 5nm SoC featuring four Cortex-A78 CPU cores clocked at 2.4GHz and four Cortex-A55 cores at 2.0 GHz.

The CPU is paired with ARM’s Mali G68 GPU clocked at 950MHz. The tablet showed up with 6GB RAM, though there have been rumors of a 4GB RAM variant. Samsung will pair the two RAM variants with 64GB and 128GB storage, respectively. The Google Play Console also confirmed a WUXGA (1920 x 1200 pixels) display for the Galaxy Tab Active 5 5G with a pixel density of 320 dpi.

It is rumored to be an 8-inch TFT LCD panel with Corning Gorilla Glass 5 protection and enhanced touch sensitivity. Since we are talking about a rugged tablet, that’s the minimum we expect from Samsung. Leaked renders have already revealed the tablet’s design. It features a user-removable metallic back plate, a sturdy body with chunky bezels and bumpers around the display, and a customizable Active Key.

Leaks have also revealed that the Galaxy Tab Active 5 5G will feature a side-mounted fingerprint scanner, a 3.5mm headphone jack, a 13MP ream camera with LED flash and 4K video recording, a 5MP selfie camera, a microSD card slot (supports up to 1TB cards), a 5,050mAh battery with 25W wired charging, Wi-Fi 6, Bluetooth v5.3, NFC, Dolby Atmos audio, Samsung Knox security, built-in S Pen support, and Android 14.

The device might launch alongside the Galaxy Xcover 7

Samsung might be planning to launch the Galaxy Tab Active 5 alongside the Galaxy Xcover 7, its next-gen rugged smartphone. We recently published exclusive official renders of the Galaxy Xcover 7, showing its design from various angles. A launch date for these devices is still missing, though. The Korean firm is currently focused on the Galaxy S24 launch event on January 17. We expect to learn more about the new rugged products after the event next week.

Samsung Galaxy Tab Active 5 5G Google Play Console


[ad_2]
Source link

Volkswagen plugs ChatGPT into its cars and SUVs

andreasc
-
0
Volkswagen plugs ChatGPT into its cars and SUVs
[ad_1]

The AI fever is on! Volkswagen, among the many, has embraced this technology. Notably, Volkswagen announced its plans to integrate OpenAI’s ChatGPT into its vehicles at CES 2024. Expectedly, this integration will start rolling out in the second quarter of 2024. Now, you might be wondering about the necessity of having a Volkswagen ChatGPT car. However, Volkswagen is prepared with an answer. They strongly believe that upcoming updates will show how useful ChatGPT can be. They say ChatGPT will make talking in the car better, give quick answers, let you chat naturally, share car information, and all of this without needing to touch anything—just by speaking.

Volkswagen’s ChatGPT integration will first come to Europe

Volkswagen’s ChatGPT integration will be available across its lineup, including in Tinguan, Passat, and Golf, as well as their electric ID series. Initially, it will come to Europe, with potential plans for a US release still under discussion.

Volkswagen is making its in-car voice assistant better by adding ChatGPT. This means people in the car can talk more easily with the car’s system. People can use this voice assistant to do things like change the temperature or ask questions. But, sometimes ChatGPT might give wrong answers so it’s good to be careful.

Volkswagen won’t force you to download new apps to use this chatbot

Volkswagen says you don’t need to make a new account or download any apps to use their chatbot. You can turn it on by saying “Hello IDA” or pressing a button on your steering wheel. OpenAI won’t see any of your driving data, as Volkswagen deletes all questions and answers right away to keep your information safe.

Most car voice assistants can only do basic things like turning on heaters. They struggle to have good conversations or understand complex directions. Sometimes, you have to repeat what you say. Big car companies often use voice assistants from big tech companies like Google and Amazon.

But Volkswagen is different. Even if some people say their ChatGPT makes mistakes, Volkswagen keeps going. While other companies are still thinking about using AI in cars, Volkswagen is moving forward. They are working with a company called Cerence to make sure ChatGPT works well in their cars.


[ad_2]
Source link

Cacti Blind SQL Injection Flaw Enables Remote Code Execution

andreasc
-
0
Cacti Blind SQL Injection Flaw Enables Remote Code Execution
[ad_1]

Cacti, the performance and fault management framework, has been discovered with a blind SQL injection vulnerability, which could reveal Cacti database contents or trigger remote code execution. 

The CVE for this vulnerability has been assigned with CVE-2023-51448, and the severity has been given as 8.8 (High).

This vulnerability existed due to insufficient sanitization of a parameter. However, Cacti has swiftly upon this vulnerability and patched this vulnerability.

Document
Free Webinar

Compounding the problem are zero-day vulnerabilities like the MOVEit SQLi, Zimbra XSS, and 300+ such vulnerabilities that get discovered each month. Delays in fixing these vulnerabilities lead to compliance issues, these delay can be minimized with a unique feature on AppTrana that helps you to get “Zero vulnerability report” within 72 hours.

 Cacti Blind SQL Injection Flaw

An authenticated threat actor with any account that has the privilege to “Settings/Utilities” could exploit this vulnerability by sending a crafted HTTP GET request to the “/managers.php” endpoint, which could result in the SQLi payload being executed.

The parameter is vulnerable to this SQL injection attack has been identified as ‘selected_graph_array.’

Vulnerable Code
Vulnerable Code (Source: GitHub)

According to the reports shared with Cyber Security News, this vulnerability arises due to the function ‘form_actions’ in the file ‘managers.php.’ In addition to this, the ‘get_nfilter_request_var’ does not perform sanitization when deserializing the provided value. 

It was also stated that the ‘cacti_unserialize’ function uses the ‘allowed_classes’ argument to remove any PHP Object Injection vectors but does not perform any sanitization on the deserialized value.

Hence, the flow of this attack would start with the attacker sending a serialized array which has the SQLi payload deserialized to an in-memory array followed by joining it into a single string and concatenating it into a raw SQL query—this query, when executed in the database, results in exposing sensitive information.

Moreover, the impact of this vulnerability may vary based on the configuration of the Cacti database, which could even result in arbitrary file reading and writing.

Furthermore, this vulnerability can also be escalated to Remote code execution, as demonstrated in another vulnerability, CVE-2023-49084. 

Users of Cacti are recommended to upgrade to the patched version to prevent this vulnerability from getting exploited by threat actors.


[ad_2]
Source link

OPPO Find X7 Ultra’s HyperTone Image Engine is coming to more phones

andreasc
-
0
OPPO Find X7 Ultra’s HyperTone Image Engine is coming to more phones
[ad_1]

OPPO unveiled its new flagship smartphone yesterday, the OPPO Find X7 Ultra. Along with it, the company also announced the so-called ‘HyperTone Image Engine’. In OPPO’s words, the HyperTone Image Engine “safeguards photo integrity, preserving highlights, shadows, and mid-tones with a natural look”. Well, it is now official, the HyperTone Image Engine is coming to more OPPO phones.

OPPO’s HyperTone Image Engine is coming to more phones

The company said that would happen, while several phones were mentioned by name. This feature is coming to the OPPO Find N3 and Find N3 Flip handsets, the company’s latest foldable offerings. In addition to those devices, the OPPO Reno11 Pro will also get it.

OPPO did mention that some features are hardware-dependent when it comes to the HyperTone Image Engine. However, the core algorithms, photography characteristics, and style will come to more devices.

The HyperTone Image Engine brings AI denoising in the RAW domain. In addition to that, it also offers individual pixel brightness correction with photon matrix technology. As a result, you should get better-looking photos overall.

This image processing can increase image clarity & reduce noise by quite a bit, amongst other things

OPPO specifically said that HyperTone Image Engine can increase image clarity by 30%. It can also reduce image noise by 60%. It is essentially the company’s brand-new image processing.

The OPPO Find N3 already has great camera performance, but this new image processing tech could push things to a whole new level. We’re not sure when is the update coming, however.

It will land on the Chinese models first, of course. The aforementioned devices do have global models, however. So, it is coming to them as well. OPPO did not really share any information regarding that, unfortunately.

The OPPO Find X7 Ultra is the first phone to debut the tech. That phone also has an outstanding hardware camera setup on the back, arguably the best we’ve seen in the market thus far. Not only does it debut the Sony LYT-900 sensor, but it also has two periscope telephoto cameras. OPPO refers to this camera group as the ‘Quad Main Camera’ setup.


[ad_2]
Source link
1...545546547...1,482Page 546 of 1,482