Why people are saying two-factor authentication isn’t perfect

When two-factor authentication was first introduced, it revolutionized device security and helped make identity theft much more difficult – at the slight cost of minor inconvenience added to logins.

But it’s not perfect, nor has it solved all of our hacking and data theft problems. Some recent news has provided more context for how hackers have been sidestepping two-factor authentication and eroding some of our trust in it.

What exactly is two-factor authentication?

Two-factor authentication adds an extra layer of security to the login process for devices and services. Previously, logins had a single factor for authentication — typically, a password, or a biometric login like a fingerprint scan or Face ID, occasionally with the addition of security questions. That provided some security, but it was far from perfect, especially with weak passwords or autofilled passwords (or if login databases are hacked and that info starts showing up on the dark web).

Two-factor authentication addresses these issues by adding a second factor, another thing a person hasto do to guarantee that it’s really them and they have authority to access. Typically, that means being sent a code via another channel, like getting a text message or email from the service, which you then have to input.

Duo Authentication Example.

Some use time-sensitive codes (TOTP, Time-Based One Time Password), and some use unique codes associated with a specific device (HOTP, HMAC-based One Time Password). Certain commercial versions may even use additional physical keys that you need to have at hand.

The security feature has become so common, you’re probably used to seeing messages along the lines of, “We’ve sent you an email with a secure code to enter, please check your spam filter if you haven’t received it.” It’s most common for new devices, and while it takes a little time, it’s a huge jump in security compared to one-factor methods. But there are some flaws.

That sounds pretty secure. What’s the problem?

A report came out recently from cybersecurity company Sophos that detailed a surprising new way that hackers are skipping over two factor authentication: cookies. Bad actors have been “cookie stealing,” which gives them access to virtually any kind of browser, web service, email account, or even file.

How do these cybercriminals get these cookies? Well, Sophos notes that the Emotet botnet is one such cookie-stealing piece of malware that targets data in Google Chrome browsers. People can also purchase stolen cookies through underground marketplaces, which was made famous in the recent EA case where login details ended up on a marketplace called Genesis. The result was 780 gigabytes of stolen data that was used to try and extort the company.

While that’s a high-profile case, the underlying method is out there, and it shows that two-factor authentication is far from a silver bullet. Beyond just cookie stealing, there are a number of other issues that have been identified over the years:

  • If a hacker has gotten hold of your username or password for a service, they may have access to your email (especially if you use the same password) or phone number. This is especially problematic for SMS/text-based two-factor authentication, because phone numbers are easy to find and can be used to copy your phone (among other tricks) and receive the texted code. It takes more work, but a determined hacker still has a clear path forward.
  • Separate apps for two-factor authentication, like Google Auth or Duo, are far more secure, but adoption rates are very low. People tend to not want to download another app just for security purposes for a single service, and organizations find it a lot easier to simply ask “Email or text?” rather than require customers to download a third-party app. In other words, the best types of two-factor authentication aren’t really being used.
  • Sometimes passwords are too easy to reset. Identity thieves can gather enough information about an account to call up customer service or find other ways to request a new password. This often circumvents any two-factor authentication involved and, when it works, it allows thieves direct access to the account.
  • Weaker forms of two-factor authentication offer little protection against nation-states. Governments have tools that can easily counter two-factor authentication, including monitoring SMS messages, coercing wireless carriers, or intercepting authentication codes in other ways. That’s not good news for those who want ways to keep their data private from more totalitarian regimes.
  • Many data theft schemes bypass two-factor authentication entirely by focusing on fooling humans instead. Just look at all the phishing attempts that pretend to be from banks, government agencies, internet providers, etc., asking for important account information. These phishing messages can look very real, and may involve something like, “We need your authentication code on our end so we can also confirm you are the account holder,” or other tricks to get codes.

Should I keep on using two-factor authentication?

Absolutely. In fact, you should go through your services and devices and enable two-factor authentication where it’s available. It offers significantly better security against problems like identity theft than a simple username and password.

Even SMS-based two-factor authentication is much better than none at all. Infact, the National Institute of Standards and Technology once recommended against using SMS in two-factor authentication, but then rolled that back the next year because, despite the flaws, it was still worth having.

When possible, choose an authentication method that’s not connected to text messages, and you’ll have a better form of security. Also, keep your passwords strong and use a password manager to generate them for logins if you can.

Security and Privacy settings open on a MacBook.

How can two-factor authentication be improved?

Moving away from SMS-based authentication is the big current project. It’s possible that two-factor authentication will transition to a handful of third-party apps like Duo, which remove many of the weaknesses associated with the process. And more high-risk fields will move into MFA, or multi-factor authentication, which adds a third requirement, like a fingerprint or additional security questions.

But the best way to remove issues with two-factor authentication is to introduce a physical, hardware-based aspect. Companies and government agencies are already starting to require that for certain access levels. In the near future, there’s a fair chance we’ll all have customized authentication cards in our wallets, ready to swipe at our devices when logging into services. It may sound weird now, but with the steep rise of cybersecurity attacks, it could end up being the most elegant solution.

Editors’ Choice

Repost: Original Source and Author Link


Why people are saying the RTX 4090 isn’t worth waiting for

Nvidia’s long-awaited RTX 4090 is expected to launch within a few months, ushering in a new era for the best graphics cards alongside AMD. We don’t have any official news on the RTX 4090 yet, but the rumor mill hasn’t been quiet.

A slew of leakers and industry insiders say Nvidia’s next-gen GPU might not be worth waiting for, especially considering the current state of GPU prices.

It draws a ton of power

Jacob Roach / Digital Trends

Leaks have shown that the RTX 4090 may be the most powerful GPU ever tested, but all that performance comes at a price: power. Rumored specs suggest the card could draw in the range of 400 watts to 450 watts, which is a massive jump over even the RTX 3090 (it draws 350W). Other rumors say it can climb as high as 600W under full load, though I imagine Nvidia will limit the power quite a bit below the 600W mark.

Regardless, power is a concern. The last generation of GPUs has been a swift kick to PC gamers to upgrade their power supply, and the RTX 4090 may prompt another upgrade. More power means more heat too, which is a big problem when gamers are already downclocking their GPUs to reduce heat and power draw.

Power demands have been climbing for high-end gaming PCs, and the next generation of GPUs is shaping up to push power limits further. Leakers and rumor brokers say that results in much higher performance, though. Some suggest we could as much as a 2.2x increase in performance over the RTX 3090, which is a much bigger generational jump than we normally see.

More performance for the sake of more performance isn’t a winning strategy for upgrading your gaming PC, though. Nvidia’s RTX 3090 and RTX 3090 Ti still shred through any modern AAA game at 4K and a smooth 60 frames per second (fps), and tools like Deep Learning Super Sampling (DLSS) and FidelityFX Super Resolution (FSR) offer ways to boost your performance without sucking up more power.

It’s massive

An alleged render of Nvidia's RTX 3090 Ti GPU.

Leaked renders of the RTX 4090 show a comically large graphics card. It’s possible these renders are fake and that Nvidia is using a different design, but a few respected leakers have gone on the record saying the card will come with a massive cooler. And you don’t have to look any further than current-gent RTX 30-series cards to assume that will be the case.

RTX 3090 and RTX 3090 Ti models usually take up three slots and can occupy as many as four. Given the high power demands that leakers are saying will come with the RTX 4090, it wouldn’t be surprising to see a four-slot reference design and maybe even five-slot designs from Nvidia’s partners.

Well-known leaker kopite7kimi said that the size may not reach five slots, but Nvidia is still increasing the size. The leaker says the RTX 4090 will be the size of an RTX 3090 Ti, the RTX 4070 will be the size of an RTX 3070 Ti, and so on.

Just like power, size doesn’t inherently make the RTX 4090 a bad choice. It’s something to consider if you’re holding out for a card, though, especially considering we’re seeing some of the lowest GPU prices in nearly two years.

Price and availability apply

A graphics card in neon lights.
Aleksandr Grechanyuk/Shutterstock

That brings up the main elephant in the room: pricing. There aren’t any rumors that the RTX 4090 will be more expensive than the RTX 3090. Some leakers are saying it’ll be priced at the same $1,500, in fact. The RTX 3090 is seeing a massive price decrease right now, though. Some models are selling for around $1,100, and a $400 savings is nothing to sneeze at. Keep in mind that third-party RTX 4090 models will sell for more than the price set by Nvidia, maybe as much as $500 more for certain high-end models.

Availability could also be an issue, but leakers haven’t said that Nvidia is worried. New generations just prompt a minor shortage of cards as the initial units sell out. If you don’t get at card a launch, be prepared to wait for a few weeks (or maybe even a month or two) to find a card in stock at list price.

Nvidia lost around $1 billion in gaming revenue in the first half of this year, so I don’t suspect we’ll see big price hikes for the RTX 4090. Next-gen GPU prices all come down to where crypto is at, though. Cryptocurrency was a major driving force behind the GPU shortage throughout 2020 and 2021, and another spike in interest could push up GPU prices.

That’s not the case right now. GPU prices are lower than they’ve been in nearly two years, and cards are readily available below list price at retailers. You can wait until October when leakers say the RTX 4090 will launch, but you could miss a great deal now and be left with a GPU that draws too much power and can barely fit inside a normal PC case.

Editors’ Choice

Repost: Original Source and Author Link


Why people are saying to buy the M1 MacBook Air over the M2

The once highly anticipated M2 MacBook Air is finally out. But despite how great the design looks, many potential buyers are instead turning to the M1 MacBook Air, a laptop that’s nearly two years old.

Apple still sells it, of course, and as plenty of reviewers and commentators have pointed out, it may prove to be the better option for many people — and there are three main reasons why.


Price is the most obvious reason people have been recommending the M1 over the M2. With the introduction of the M2 MacBook Air, Apple created a family of MacBook Air models running its proprietary chips. The M1 MacBook Air is the absolute base model of this family, starting at $1,000. In comparison, the M2 MacBook Air is the base model of the M2 line starting at $1,200.

While not that pricey, it is enough to make a savvy consumer pause if a cheaper model is available. Of course, there are lots of upgrades involved in the $200 premium of the M1 MacBook Air over the M2 MacBook Air. That includes a thinner chassis, a MagSafe 3 charging port, thinner bezels, faster CPU performance, brighter screen, better speakers, and a higher resolution webcam. That’s a lot.

But because of how good the M1 already was, that $200 difference is still probably enough to deter many potential buyers who are just looking for a basic, fast laptop — whether that’s for college or work. In the end, it’s the cheapest MacBook you can buy, and that makes it undeniably attractive.

An Apple Store product page with the prices of the M1 and M2 MacBook Air next to each other.

The M2 MacBook Air base model is essentially the mid-tier device in the new MacBook Air family, and the next M2 MacBook Air configuration is the high-end model at $1,500. Devices can get pricier still with unique configurations at any tier, but the M1 MacBook Air will still give consumers the most bang for their buck at any hardware style.

While Apple continues to struggle with chip shortages, your best bet access-wise is the M1 MacBook Air. The company faced shipping delays with the M2 MacBook Air immediately at pre-order. Purchasing suggestions for those looking to buy Apple products earlier in the year were to buy the M1 MacBook Air for the quickest shipments. Overall, being cost-effective and easy to buy for several years, both online and in stores has been a plus for this laptop.

SSD and performance

The motherboard of the M2 MacBook Air is revealed in a YouTube teardown.

The M1 and M2 MacBook Air models both feature the same SSD storage options in terms of capacity. Both laptops are available in 256GB base models, which can be configured to 512GB, 1TB, and 2TB of additional storage.

Apple set out to make the base model M2 MacBook Air look like a viable upgrade with its M2 chip and eight-core GPU; however, the truth about the device quickly unfolded as reviewers and early adopters harshly critiqued the new laptop.

Many quickly figured out that Apple used its top-tier M2 MacBook Air model when demonstrating the performance power of the laptop during its WWDC unveiling in June and when sending out official review models. When Apple said the M2 MacBook Air would have an 18% increase in graphics performance over the M1, the company was talking about the higher-end model, which features a 10-core GPU and 512GB SSD.

Meanwhile, teardowns uncovered the base model M2 MacBook Air was missing features that could impact performance. It included a single NAND chip, instead of two flash chips, like the M1 MacBook Air, which provides faster performance. YouTuber Max Tech discovered in benchmarks that the base model M2 MacBook Air actually had 50% slower read/write speeds than the M1. The lower storage performance is not something the M1 MacBook Air suffers from, meaning there are situations where the M1 model is a better performer, such as in file transfers.

This, coupled with a thinner design and a lot of thermal paste issues, made the M2 MacBook Air susceptible to overheating. These performance flubs have made the M1 MacBook Air look like the favorable base model laptop, especially for the price.

No notch

Apple MacBook Air M1 open, on a table.
Andy Boxall/Digital Trends

The M2 MacBook Air has been released with a larger 13.6-inch display and thinner bezels than the 13.3-inch M1 MacBook Air. The M2 display is also slightly brighter at 500 nits versus 400 nits on the M1 display. However, the M2 MacBook Air has an additional feature: the controversial notch design, which houses the laptop’s 1080p webcam.

You might enjoy the new look of the screen, but when given the option of thinner bezels or a notch, it’s a bit less of an obvious choice. The inclusion of an important feature like Face ID may have made the notch easier to stomach, but that’s still a feature we’re waiting on. As it is, the notch will feel like an exciting new design element of the MacBook Air, and more like a feature you’ll just have to learn to live with. For a laptop that costs more, some may find it better to just stick with the M1.

Editors’ Choice

Repost: Original Source and Author Link


People spent much less time watching gaming streams this spring, report says

The number of hours streamed and watched across , and have dropped significantly over the last year, according to the latest Streamlabs and Stream Hatchet on the landscape of livestreaming. Between April and June, streamers on the three platforms were live for 273 million hours. That’s down 19.4 percent from Q2 2021 and 12 percent from the previous quarter.

Viewers tuned in to streams for 7.36 billion hours across the three platforms last quarter. That’s a drop of 18.1 percent year over year (viewership was at 8.99 billion hours in Q2 2021) and 8.4 percent from the previous quarter. The slowdown for all three platforms could be a case of people spending more time outside than they did last year for pandemic-related reasons.

Twitch is still by far the biggest player among the three platforms, with 76.7 percent of market share in terms of hours watched (5.64 billion) and 92.7 percent of hours streamed (204.2 million). Those figures dropped by 13.4 percent and 16 percent from Q2 2021. The number of unique channels streaming on the platform dropped by nearly 2 million to 9.6 million as well.

However, Twitch’s Just Chatting category continues to go from strength to strength. Hours watched there actually grew by 2.2 percent from the previous quarter, giving the category its highest ever viewership. The most-watched categories after that were (465 million hours) and (464 million).

YouTube Gaming viewership actually remained steady from the previous quarter, though it dropped 13.1 percent from Q2 2021 to 1.13 billion hours. The total hours streamed dropped by 9.6 percent year over year to 8.05 million.

Facebook Gaming suffered a bigger setback, per the report, despite Meta’s efforts to court creators. The number of hours watched fell by a whopping 51 percent from a year ago to 580 million. There was an even bigger drop in terms of hours streamed, from 20.8 million in Q2 2021 to 7.9 million last quarter — a decline of 62 percent.

Perhaps we’ll soon start seeing some of those numbers creep up again, though. With a recession looming, folks may spend more time indoors again, tuning back into streamers they enjoyed watching during the first 18 months or so after COVID-19 took hold.

All products recommended by Engadget are selected by our editorial team, independent of our parent company. Some of our stories include affiliate links. If you buy something through one of these links, we may earn an affiliate commission.

Repost: Original Source and Author Link


Why people are saying to avoid the $1,199 M2 MacBook Air

Apple’s latest MacBook Air with the new M2 chip has been controversial, to say the least. The new MacBook Air features a refreshed design and is the successor to the M1, Apple’s new M2 system-on-a-chip.

But since the new M2 MacBook Air dropped, tech reviewers, experts, and regular people are saying you should avoid the base configuration of this new machine due to some pretty substantial performance issues. For a more affordable laptop, this is a frustrating bind that potential MacBook Air buyers are being put in.

For starters, the $1,199 base configuration has an eight-core GPU, while the upgraded version has a full 10-core GPU. This isn’t what Apple told us at WWDC when it showed off the new Air. Apple promised an 18% increase in graphical performance over the M1. What the company didn’t say, however, was the model it showed off was the more expensive upgraded M2 MacBook Air, which retails for $1,499. The model with an eight-core GPU will have a much smaller performance advantage compared to the seven-core M1 MacBook Air.

Then there’s the issue of the slow SSD. The base model comes with a measly 256GB of SSD storage, which itself is an insult to consumers for nearly $1,200. But the problem is bigger than storage space.

The SSD features a single NAND chip instead of two, as is the norm. This is reportedly producing far slower read/write speeds than the 512GB model. In fact, the base model M2 MacBook Air’s storage is a whopping 50% slower than 2020’s M1 MacBook Air, which featured two flash chips.

Just look at these benchmarks posted by YouTuber Max Tech:

  • 2020 M1 chip read/write speed: 2900/2215
  • 2022 M2 chip read/write speed: 1446/1463

But wait, it gets worse. The combination of the slow SSD and the 8GB of shared memory actually bottleneck performance so hard that according to Max Tech, there can even be tasks such as file transfers in which it’s even slower than the M1 MacBook Air, which doesn’t have this same deficiency. Yikes. The issue was first discovered in the M2 MacBook Pro 13-inch, but they apply just the same to the MacBook Air.

This wasn’t something we were able to verify ourselves at DT, as our review unit came with the 512GB SSD, which was plenty zippy.

The screen of the MacBook Air on a table.

Finally, there’s no fast charger. Apple includes a basic 30-watt charging brick with this model, while the upgraded version comes with a 35-watt dual charger. It’s not much faster, but at least it’s an improvement. A $300 dollar improvement, though? You can easily buy a 67-watt fast charger for under $30 from the Apple Store. But you shouldn’t have to.

Don’t take this the wrong way. We love the new design language Apple went with for the M2 MacBook Air. The boxy, industrial minimalism look brings it in line with the rest of Apple’s devices. The M2 chip also holds so much potential for computing in general. Apple Silicon is changing the face of the computer world — there’s no question about it.

But Apple seems to be holding it back with this nerfed base model M2 MacBook Air, and almost everybody agrees. You’ll probably want to upgrade to the better configurations or opt for the M1 MacBook Air.

Editors’ Choice

Repost: Original Source and Author Link


Google is using AI to better detect searches from people in crisis

In a personal crisis, many people turn to an impersonal source of support: Google. Every day, the company fields searches on topics like suicide, sexual assault, and domestic abuse. But Google wants to do more to direct people to the information they need, and says new AI techniques that better parse the complexities of language are helping.

Specifically, Google is integrating its latest machine learning model, MUM, into its search engine to “more accurately detect a wider range of personal crisis searches.” The company unveiled MUM at its IO conference last year, and has since used it to augment search with features that try to answer questions connected to the original search.

In this case, MUM will be able to spot search queries related to difficult personal situations that earlier search tools could not, says Anne Merritt, a Google product manager for health and information quality.

“MUM is able to help us understand longer or more complex queries like ‘why did he attack me when i said i dont love him,’” Merrit told The Verge. “It may be obvious to humans that this query is about domestic violence, but long, natural-language queries like these are difficult for our systems to understand without advanced AI.”

Other examples of queries that MUM can react to include “most common ways suicide is completed” (a search Merrit says earlier systems “may have previously understood as information seeking”) and “Sydney suicide hot spots” (where, again, earlier responses would have likely returned travel information — ignoring the mention of “suicide” in favor of the more popular query for “hot spots”). When Google detects such crisis searches, it responds with an information box telling users “Help is available,” usually accompanied by a phone number or website for a mental health charity like Samaritans.

In addition to using MUM to respond to personal crises, Google says it’s also using an older AI language model, BERT, to better identify searches looking for explicit content like pornography. By leveraging BERT, Google says it’s “reduced unexpected shocking results by 30%” year-on-year. However, the company was unable to share absolute figures for how many “shocking results” its users come across on average, so while this is a comparative improvement, it gives no indication of how big or small the problem actually is.

Google is keen to tell you that AI is helping the company improve its search products — especially at a time when there’s a building narrative that “Google search is dying.” But integrating this technology comes with its downsides, too.

Many AI experts warn that Google’s increasing use of machine learning language models could surface new problems for the company, like introducing biases and misinformation into search results. AI systems are also opaque, offering engineers restricted insight into how they come to certain conclusions.

For example, when we asked Google how it verifies in advance which search terms identified by MUM are associated with personal crises, its reps were either unwilling or unable to answer. The company says it rigorously tests changes to its search products using human evaluators, but that’s not the same as knowing in advance how your AI system will respond to certain queries. For Google, though, such trade-offs are apparently worth it.

Repost: Original Source and Author Link


Meta expects a billion people in the metaverse by 2030

Meta believes that a billion people will be participating in the metaverse within the next decade, despite the concept feeling very nebulous at the moment.

CEO Mark Zuckerberg spoke with CNBC’s Jim Cramer on a recent broadcast of Mad Money and went on to say that purchases of metaverse digital content would bring in hundreds of billions of dollars for the company by 2030. This would quickly reverse the growing deficit of Meta’s Reality Labs, which has already invested billions into researching and developing VR and AR hardware and software.

Currently, this sounds like a stretch given that only a small percentage of the population owns virtual reality hardware and few dedicated augmented reality devices have been released from major manufacturers. Apple and Google have each developed AR solutions for smartphones and Meta has admitted that the metaverse won’t require special hardware in order to access it.

Any modern computer, tablet, or smartphone has sufficient performance to display virtual content, however, the fully immersive experience is available only when wearing a head-mounted display, whether that takes the form of a VR headset or AR glasses.

According to Cramer, Meta is not taking a cut from creators initially, while planning to continue to invest heavily into hardware and software infrastructure for the metaverse. Meta realizes it can’t build an entire world by itself and needs the innovation of creators and the draw of influencers to make the platform take off in the way Facebook and Instagram have.

Zuckerberg explained that Meta’s playbook has always been to build services that fill a need and grow the platform to a billion or more users before monetizing it. That means the next 5 to 10 years might be a rare opportunity for businesses and consumers to take advantage of a low-cost metaverse experience before Meta begins to demand a share. Just as Facebook was once ad-free, the early metaverse might be blissfully clear from distractions.

This isn’t exclusively Meta’s strategy, but the growth method employed by most internet-based companies. Focusing on growth first and money later has become standard practice. In the future, a balancing act will be required to make enough money to fund services while keeping the metaverse affordable enough to retain users.

While Meta might not get a billion people to strap on a VR headset by 2030, there’s little doubt that the metaverse will become an active area of growth. It should interest enough VR, AR, smartphone, tablet, and computer owners to be self-sustaining within a few years and could actually explode to reach a billion people by 2030.

Editors’ Choice

Repost: Original Source and Author Link


Deep North, which uses AI to track people from camera footage, raises $16.7M

Join gaming leaders online at GamesBeat Summit Next this upcoming November 9-10. Learn more about what comes next. 

Deep North, a Foster City, California-based startup applying computer vision to security camera footage, today announced that it raised $16.7 million in a series A-1 round. Led by Celesta Capital and Yobi Partners, with participation from Conviction Investment Partners, Deep North plans to use the funds to make hires and expand its services “at scale,” according to CEO Rohan Sanil.

Deep North, previously known as Vmaxx, claims its platform can help brick-and-mortar retailers “embrace digital” and protect against COVID-19 by retrofitting security systems to track purchases and ensure compliance with masking rules. But the company’s system, which relies on algorithms with potential flaws, raises concerns about both privacy and bias.

“Even before a global pandemic forced retailers to close their doors … businesses were struggling to compete with a rapidly growing online consumer base,” Sanil said in a statement. “As stores open again, retailers must embrace creative digital solutions with data driven, outcome-based computer vision and AI solutions, to better compete with online retailers and, at the same time, accommodate COVID-safe practices.”

AI-powered monitoring

Deep North was founded in 2016 by Sanil and Jinjun Wang, an expert in multimedia signal processing, pattern recognition, computer vision, and analytics. Wang — now a professor at Xi’an Jiaotong University in Xi’an, China — was previously a research scientist at NEC before joining Epson’s R&D division as a member of the senior technical staff. Sanil founded a number of companies prior to Deep North, including Akirra Media Systems, where Wang was once employed as a research scientist.

“In 2016, I pioneered object detection technology to help drive targeted advertising from online videos. When a major brand saw this, they challenged m e to create a means of identifying, analyzing, and sorting objects captured on their security video cameras in their theme parks,” Sanil told VentureBeat via email. “My exploration inspired development that would unlock the potential of installed CCTV and security video cameras within the customer’s physical environment and apply object detection and analysis in any form of video.”

After opening offices in China and Sweden and rebranding in 2018, Deep North expanded the availability of its computer vision and video analytics products, which offer object and people detection capabilities. The company says its real-time, AI-powered and hardware-agnostic software can understand customers’ preferences, actions, interactions, and reactions “in virtually any physical setting” across “a variety of markets,” including retailers, grocers, airports, drive-thrus, shopping malls, restaurants, and events.

Deep North says that retailers, malls, and restaurants in particular can use its solution to analyze customer “hotspots,” seating, occupancy, dwell times, gaze direction, and wait times, leveraging these insights to figure out where to assign store associates or kitchen staff. Stores can predict conversion by correlating tracking data with the time of day, location, marketing events, weather, and more, while shopping centers can draw on tenant statistics to understand trends and identify “synergies” between tenants, optimizing for store placement and cross-tenant promotions.

Deep North

“Our algorithms are trained to detect objects in motion and generate rich metadata about physical environments such as engagement, pathing, and dwelling. Our inference pipeline brings together camera feeds and algorithms for real-time processing,” Deep North explains on its website. “[We] can deploy both via cloud and on-premise and go live within a matter of hours. Our scalable GPU edge appliance enables businesses to bring data processing directly to their environments and convert their property into a digital AI property. Video assets never leave the premise, ensuring the highest level of security and privacy.”

Beyond these solutions, Deep North developed products for particular use cases like social distancing and sanitation. The company offers products that monitor for hand-washing and estimate wait times at airport check-in counters, for example, as well as detect the presence of masks and track the status of maintenance workers on tarmacs.

“With Deep North’s mask detection capability, retailers can easily monitor large crowds and receive real-time alerts,” Deep North explains about its social distancing products. “In addition, Deep North … monitors schedules and coverage of sanitization measures as well as the total time taken for each cleaning activity … Using Deep North’s extensive data, [malls can] create tenant compliance scorecards to benchmark efforts, track overall progress, course-correct as necessary. [They] can also ensure occupancy limits are adhered to across several properties, both locally and region-wide, by monitoring real-time occupancy on our dashboard and mobile apps.”

Bias concerns

Like most computer vision systems, Deep North’s were trained on datasets of images and videos showing examples of people, places, and things. Poor representation within these datasets can result in harm — particularly given that the AI field generally lacks clear descriptions of bias.

Previous research has found that ImageNet and Open Images — two large, publicly available image datasets — are U.S.- and Euro-centric, encoding humanlike biases about race, ethnicity, gender, weight, and more. Models trained on these datasets perform worse on images from Global South countries. For example, images of grooms are classified with lower accuracy when they come from Ethiopia and Pakistan, compared to images of grooms from the United States. And because of how images of words like “wedding” or “spices” are presented in distinctly different cultures, object recognition systems can fail to classify many of these objects when they come from the Global South.

Bias can arise from other sources, like differences in the sun path between the northern and southern hemispheres and variations in background scenery. Studies show that even differences between camera models — e.g., resolution and aspect ratio — can cause an algorithm to be less effective in classifying the objects it was trained to detect.

Tech companies have historically deployed flawed models into production. ST Technologies’ facial recognition and weapon-detecting platform was found to misidentify black children at a higher rate and frequently mistook broom handles for guns. Meanwhile, Walmart’s AI- and camera-based anti-shoplifting technology, which is provided by Everseen, came under scrutiny last May over its reportedly poor detection rates.

Deep North doesn’t disclose on its website how it trained its computer vision algorithms, including whether it used synthetic data (which has its own flaws) to supplement real-world datasets. The company also declines to say to what extent it takes into account accessibility and users with major mobility issues.

In an email, Sanil claimed that Deep North “has one of the largest training datasets in the world,” derived from real-world deployments and scenarios. “Our human object detection and analysis algorithms have been trained with more than 130 million detections, thousands of camera feeds, and various environmental conditions while providing accurate insights for our customers,” he said. “Our automated and semi-supervised training methodology helps us build new machine learning models rapidly, with the least amount of training data and human intervention.”

In a follow-up email, Sanil added: “Our platform detects humans, including those with unique gaits, and those that use mobility aids and assistive devices. We don’t do any biometric analysis, and therefore there is no resulting bias in our system … In the simplest terms, the platform interprets everything as an object whether it’s a human or a shopping cart or a vehicle. We provide object counts entering or exiting a location. Our object counting and reporting is not influenced by specific characteristics.” He continued: “We have a large set of labeled data. For new data to be labeled, we need to classify some of the unlabeled data using the labeled information set. With the semi-supervised process we can now expedite the labeling process for new datasets. This saves time and cost for us. We don’t need annotators, or expensive and slow processes.”

Privacy and controversy

While the purported goal of products like Deep North’s are health, safety, and analytics, the technology could be coopted for other, less humanitarian intents. Many privacy experts worry that they’ll normalize greater levels of surveillance, capturing data about workers’ movements and allowing managers to chastise employees in the name of productivity.

Deep North is no stranger to controversy, having reportedly worked with school districts and universities in Texas, Florida, Massachusetts, and California to pilot a security system that uses AI and cameras to detect threats. Deep North claims that the system, which it has since discontinued, worked with cameras with resolutions as low as 320p and could interpret people’s behavior while identifying objects like unattended bags and potential weapons.

Deep North is also testing systems in partnership with the U.S. Transportation Security Administration, which furnished it with a grant last March. The company received close to $200,000 in funding to provide metrics like passenger throughput, social distancing compliance, agent interactions, and bottleneck zones as well as reporting of unattended baggage, movement in the wrong direction, or occupying restricted areas.

“We are humbled and excited to be able to apply our innovations to help TSA realize its vision of improving passenger experience and safety throughout the airport,” Sanil said in a statement. “We are committed to providing the U.S. Department of Homeland Security and other government entities with the best AI technologies to build a safer and better homeland through continued investment and innovation.”

Deep North admitted in an interview with Swedish publication Breakit that it offers facial characterization services to some customers to estimate age range. And on its website, the startup touts its technologies’ ability to personalize marketing materials depending on a person’s demographics, like gender. But Deep North is adamant that its internal protections prevent it from ascertaining the identity of any person captured via on-camera footage.

“We have no capability to link the metadata to any single individual. Further, Deep North does not capture personally identifiable information (PII) and was developed to govern and preserve the integrity of each and every individual by the highest possible standards of anonymization,” Sanil told TechCrunch in March 2020. “Deep North does not retain any PII whatsoever, and only stores derived metadata that produces metrics such as number of entries, number of exits, etc. Deep North strives to stay compliant with all existing privacy policies including GDPR and the California Consumer Privacy Act.”

To date, 47-employee Deep North has raised $42.3 million in venture capital.


VentureBeat’s mission is to be a digital town square for technical decision-makers to gain knowledge about transformative technology and transact.

Our site delivers essential information on data technologies and strategies to guide you as you lead your organizations. We invite you to become a member of our community, to access:

  • up-to-date information on the subjects of interest to you
  • our newsletters
  • gated thought-leader content and discounted access to our prized events, such as Transform 2021: Learn More
  • networking features, and more

Become a member

Repost: Original Source and Author Link


T-Mobile investigating report of customer data breach that reportedly involves 100 million people

T-Mobile confirmed Sunday that it’s looking into an online forum post that claims to be selling a large trove of its customers’ sensitive data. Motherboard reported that it was in contact with the seller of the data, who said they had taken data from T-Mobile’s servers that included Social Security numbers, names, addresses, and driver license information related to more than 100 million people. After reviewing samples of the data, Motherboard reported it appeared authentic.

“We are aware of claims made in an underground forum and have been actively investigating their validity,” a T-Mobile spokesperson said in an email to The Verge. “We do not have any additional information to share at this time.”

It’s not clear when the data may have been accessed, but T-Mobile has been the target of several data breaches in the last few years, most recently in December 2020. During that incident, call-related information and phone numbers for some of its customers may have been exposed, but the company said at the time that it did not include more sensitive info such as names or Social Security numbers.

In 2018, hackers accessed personal information for roughly 2 million T-Mobile customers that included names, addresses, and account numbers, and in 2019, some of T-Mobile’s prepaid customers were affected by a breach that also accessed names, addresses, and account numbers.

A March 2020 breach exposed some T-Mobile customers’ financial information, Social Security numbers, and other account information.

Repost: Original Source and Author Link


8 Spooky Halloween Games for People Who Don’t Like Horror

October is spooky season, where vampires, zombies, and cat people come out to play. A lot of people live for Halloween and look forward to the horrorfest every year. Unfortunately, I’m a weenie. I don’t like horror — never have, never will.

Still, I want to play some fitting games to get in the spirit of the season, so I’ve been finding some less scary options. They’re games that only kind of give you the chills or take place in pretty autumn settings, perfect for channeling your inner pumpkin spice. They’re also critically acclaimed games that should at least give you something to talk to your friends about, even if you end up not liking them. Best of all, they won’t keep you up at night or make the game barely playable because you’re afraid to turn the corner.

Here are a few picks for people who don’t like horror. Alternatively, if you’re not a weenie, you can instead look at our recommendations for the best horror games of all time.

Luigi’s Mansion 3

Luigi’s Mansion 3 replaces the typical haunted mansion with a haunted hotel. Luigi, Mario, and Princess Peach accept an invite to a luxurious resort, not realizing that it’s actually a bed of supernatural activity. Our hero Luigi wakes up from a nap in his hotel room to find everyone gone and the hotel overrun by ghosts. He, along with his trusty ghost dog Polterpup, must clean up the hotel’s ghostly infestation and save his brother and friends from the vengeful King Boo.

Luigi’s Mansion fits the Halloween theme with its haunted vibe — think Casper the Friendly Ghost. It’s not meant to terrify players as much as it is to tell a story that happens to have ghosts. Plus, its cartoonish graphics lighten the blow when it comes to scares. It’s easier to downplay any spookiness when a ball-nosed cartoon plumber is running around vacuuming blob-bodied ghosts. No eerily realistic rotting skin, chilling background music, or blood baths to be seen in these haunted halls.

Luigi’s Mansion 3 is available on the Nintendo Switch. Other Luigi’s Mansion games are fine picks for Halloween too, but Luigi’s Mansion 3 is the most modern one. As Digital Trends’ review put it, “Exploring all the different rooms with all the carefully added details and clever ghost encounters has a greater impact than it did in previous games.”

Little Nightmares 2

mono and six together

Little Nightmares 2 takes place in a dark, cluttered universe eerily similar to our own. Mono, a young boy wearing a paper bag over his head, finds himself trapped in this world that’s been distorted by a mysterious signal tower. He meets Six, the little girl wearing a yellow raincoat from Little Nightmares, and the two work together to uncover the secrets of the tower and save Six from her fate.

This prequel to Little Nightmares scares players in an unsettling, bubbling at the pit of your stomach kind of way. Its shadowy, bleak setting and silent protagonists moving about a dangerous area of spooky residents stir up a sense of unease. I kept it in this list because of how its suspense-filled story convinced me to continue through cryptic corridors, even though I felt it was scarier than what I was used to.

Jump into this realm of nightmares on PC, PS4, PS5, Xbox One, Xbox Series X/S, Nintendo Switch, and Google Stadia. Alternatively, you can play Very Little Nightmares on mobile devices, which changes gameplay and graphics for a more easygoing spinoff adventure.

Famicom Detective Club

famicom detective club protagonists

Famicom Detective Club is a series, not just one title. Famicom Detective Club: The Missing Heir introduces an amnesiac protagonist who discovers that he’s a detective in the middle of solving a murder related to the wealthy Ayashiro family. On the other hand, Famicom Detective Club: The Girl Who Stands Behind stars the same protagonist before the events of Missing Heir. He investigates the murder of a schoolgirl alongside her best friend and leader of the Detective Club, Ayumi Tachibana.

Both games count as murder mystery visual novels. They aren’t made to be horror in a way that invokes a sense of creeping unease like some other titles on this list. However, they still involve murder and dead people in a way that might be entertaining for a late evening playthrough. The murders are also related to some urban legend ghost stories, which match the Halloween spirit. Overall, it checks off most elements of a scary story while keeping it light.

Both Famicom Detective Club games are available on the Nintendo Switch. You can buy one first to try out the games, or buy the entire bundle upfront for slightly cheaper than it would cost to separately buy each one.

What Remains of Edith Finch

looking at finch mansion through forest

What Remains of Edith Finch takes place through the eyes of Edith Finch, the last surviving member of the Finch family. Edith explores the abandoned Finch mansion to find out why she’s the only one left. It’s essentially an anthology of short stories about each Finch family member. You play through each Finch’s life through various interactive means until their untimely deaths.

Edith Finch has a rainy night vibe to it. Edith doesn’t seem to be in any immediate, bloody danger, but she is investigating her spooky family curse. The story explores themes of what people leave behind, but in an unnerving way that reminds us life is fleeting and our actions have consequences. As Creative Director Ian Dallas told Digital Trends, “What we’re really interested in is exploring a moment that feels very beautiful, but also a little unsettling.”

Find out what exactly remains of Edith Finch on PC, PlayStation 4, Xbox One, Nintendo Switch, and iOS.

Lost in Random

Lost in Random tells the story of one sister’s quest to save their sibling from a twisted fairytale. In the Kingdom of Random, children roll a magical die on their twelfth birthday to decide where they live for the rest of their lives. Odd rolls a six, which should mean a life of luxury in the Queen’s Castle. However, one year later, her sister Even receives a signal that indicates Odd might be in danger. Even meets a sentient die named Dicey and the two fight through different districts to save Odd.

It’s a Tim Burton-esque adventure with spindly 3D characters, dreamy settings, and mean-looking monsters. Lost in Random might have a spooky premise with the kidnapped sister and all, but it’s also a heartwarming tale without putting it in outright nightmare territory. Even the promotional materials like the trailer scream storybook come to life. Digital Trends’ reviewer mentioned some cons, but it succeeded in entertaining them with well-developed (if long-winded) characters and an intriguing mystery.

Lost in Random is on the Nintendo Switch, Xbox Series X/S, Microsoft Windows, PlayStation 5, Xbox One, PS4, and PS5.

Night in the Woods

Night in the Woods stars college dropout Mae Benson in her return to her rundown hometown of Possum Springs. Players help Mae cope with her feelings of aimlessness while uncovering something sinister brewing behind the suburban normalcy of the Western Pennsylvania-based town. It’s a hybrid genre adventure game that’s sure to entertain with its variety of mini-games and humorous, thoughtful dialogue.

It takes place in the fall, but that’s not the only reason why it’s a Halloween game. There’s a murder mystery subplot underneath this coming-of-age story. Play through the events from Halloween (or the game’s version of it) all the way toward the winter to reconnect with old friends and uncover the shady happenings that might have to do with missing people in Possum Springs.

This game might be for you if you’re looking for a young adult novel in the form of an indie hybrid adventure game with ghostly undertones. It’s available on basically every gaming platform now including PC, PS4, Xbox One, Nintendo Switch, and mobile devices. It’s definitely one of the better indie games still worth playing in 2021 and beyond.

Doki Doki Literature Club

doki doki literature club characters

Doki Doki Literature Club seems like an innocent high school dating sim, but it’s actually a psychological horror game that subverts the genre. You play as a faceless protagonist who has the option of picking between three girls: Sayori, your cheerful childhood friend, Yuri, the shy beauty with a possessive side, and Natsuki, the small feisty girl with a temper. There’s also Monika, the non-romanceable club president.

It starts off as a fairly standard sim where the player composes poems with words that represent each girl to strengthen their bonds with them. Then, everything changes when a certain cataclysmic event corrupts the entire game. Players then witness a darkening narrative with each scene until the big reveal at “the end.” This so-called sim might be for you if you like anime tropes, philosophical discussions, and creepypasta.

Doki Doki Literature Club is one of the best free-to-play games you can get on PC, PS4, and Xbox One. It also has an expansion called Doki Doki Literature Club Plus!, which includes additional content. In addition to the aforementioned platforms, the expansion is also available on the Nintendo Switch.


inside house in oxenfree

Oxenfree starts as what seems like a typical coming-of-age story before the main characters discover a ghostly rift. You play as a teenage girl named Alex, who travels to Edwards Island with her friend Ren and new stepbrother Jonas to meet up for a weekend trip. There, they meet Clarissa, the former girlfriend of Alex’s dead brother, and Nona, Clarissa’s best friend. But, just as these friends start to explore the abandoned island, their weekend getaway shifts into something spooky.

Oxenfree relies on Alex’s decisions to drive the narrative to one of the multiple endings. Players uncover Edwards Island’s dark past and determine what ultimately happens to this band of friends. Decisions can get complicated, especially with the supernatural elements like time travel, pocket dimensions, and ghosts in the story. It’s never really the bloody kind of horror, but jump scares and suspenseful moments can get a rise out of players.

Oxenfree is available on PC, PS4, PS5, Xbox One, Xbox Series X/S, and Nintendo Switch. Its sequel doesn’t come out until 2022, so you have plenty of time to catch up on the original game on your platform of choice.

Editors’ Choice

Repost: Original Source and Author Link