Connect with us

Technology

Smarter And More Practical ‘Google AI’ Techs?

Published

on

At the Google I/O, a lot of interesting innovations are coming for Google’s artificial intelligence technology that’s supposed to make it a lot more smarter and more practical in day-to-day use.

As Rishi Chandra, the vice president of product at Google says on a report, “Every decade, there’s a big computing shift. Twenty years ago, it was the transition to the web. Ten years ago or 12 years ago, it was the transition to mobile. And now, we think we’re in the third stage, which is AI or ambient computing.”

Certainly, the influx of AI-driven technology is now starting to take over the tech industry and everyone’s scramming to develop their own. We’re here to break it down and tell you all the fun facts as Google heads on to the future of AI.

Google Assistant

Throughout recent years, there has been promises over voice-driven AI tech that’s meant to make life easier by just saying a few trigger words. During an on-stage demo at the Google I/O, saying the words “Hey, Google” apparently becomes more useful and are now building on the foundation of what voice assistants can start doing.

As announced, the Google Assistant in the future will be able to perform tasks way beyond the superficial. For easy digestion, the Google Assistant will help you do tasks without the need to direct it. In further months, it won’t only be useful to do simple searches but will also help you in composing a message to a friend including putting in a subject line, searching for a picture to attach, and sending it off, among others.

The Google Assistant is also supposed to become faster through advanced AI processing. Google is able to narrow down AI models used to listen and interpret speech locally without the need to wait for data to be processed from remote servers.

However, the new Google Assistant won’t be available just yet. It is said to launch later in the year starting with Pixel phones.

Driving Mode

In supplement to the Google Assistant, it will also be integrated with its own ‘Driving Mode’ feature. Although the Google Assistant is already available in Android Auto and Maps, it will now also be available in Waze.

By simply saying “Hey, Google. Let’s drive,” it will prompt features about driving-relevant activities, top contacts, and more personal recommendations on any Android device with the Google Assistant.

With Google Assistant’s driving mode, you will be able to access personalized recommendations. For example, you have a reservation at a restaurant or a planned trip on the database, the Google Assistant will set up the fastest and the safest route to get you to your location.

Duplex

Duplex is another feature that involves the Google Assistant. Last year, Google introduced Duplex as a voice assistant that will be able to make calls and transactions on your behalf based on the information it has on you. It also has an impressive human-like AI voice to do it.

This year, Google says that it’s expanding Duplex even with web-based transactions such as chats and etc. It will be able to make you car rental reservations, movie ticket purchases or restaurant breservations. It will base information by pulling up your travel record and will auto-fill required information like your name, address, and arrival and departure time.

Google says that you will be confirmed after the transaction regarding how it went. It will indicate of a transaction was not successful due to unavailability, walk-ins only, etc.

At the moment, Google says that the Duplex is more of a preview for what is to come and have not said a concrete date on when we actually get to see and experience the tech.

Google Lens

Google Lens is a Google app that helps you comprehend information from a restaurant menu or a piece of paper just by taking a picture. This year, Google says that it can also search for photos of the dish from the menu by pulling up image results online based on information on your Google Maps to let you have an idea what it actually looks like before ordering.

There’s also a calculator function that should come in handy when paying the bill. By simply taking a photo of the receipt, Google Lens will be able to help you split the bill and add a tip.

Moreover, the Google Lens app will help put images to life. By hovering the app over Google-partnered magazines like Bon Appétit Magazine, images will start to animate and show you recipes and the like.

Lastly, the Google Lens translation tech is also getting an update by integrating it with Google Translate. Simply taking a photo of the text will prompt Google Lens to translate it. There’s also a ‘Listen’ button you can tap that will read the text in the translated version all the while highlighting the word/s it is translating.

A communications graduate writing about the latest trends and news!

Click to comment

Leave a Reply

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

Technology

Facebook May Be Tracking Your Online Photos Too

While no active hacking is happening, security experts believe that the discovery carries heavy implications.

Published

on

Photo by Glen Carrie on Unsplash

It appears that Facebook is not only tracking people’s porn-watching behavior online but also the photos they download. A researcher has found Facebook trackers in several images that he downloaded that could potentially point out to one thing: Facebook is tracking everyone.

But this shouldn’t surprise anyone. Facebook isn’t necessarily the supreme authority in data safety and cybersecurity. But the discovery made by Edin Jusupovic proves that Facebook’s tracking mission does not end in its platform.

On Twitter, Jusupovic said that Facebook has embedded “hidden tracking codes” in photos that people download. He noticed a structural abnormality when looking at a hex dump of an image file from an unknown origin only to discover it contained a special IPTC instruction.

IPTC (International Press Telecommunications Council) is an organization that sets publishing standards, including image metadata.

According to the researcher, what he found, where facebook secretly embeds tracking codes in photos only, was ““shocking level of tracking,” adding that “the take from this is that they can potentially track photos outside of their own platform with a disturbing level of precision about who originally uploaded the photo (and probably so much more).”

Yesterday, Z6Mag published a report that a group of researchers has found hidden tracking codes in different porn sites online that can be traced back to Facebook, Google, and Oracle. Researchers from Microsoft, Carnegie Mellon, and the University of Pennsylvania scanned 22,484 porn sites and found that 93% of the sade porn sites include hidden codes from the three tech giants.

Related: Study Finds Google And Facebook Trackers In Porn Sites

“Our analysis of 22,484 pornography websites indicated that 93% leak user data to a third party. Tracking on these sites are highly concentrated by a handful of major companies, which we identify. We successfully extracted privacy policies for 3,856 sites, 17% of the total. The policies were written such that one might need a two-year college education to understand them. Our content analysis of the sample’s domains indicated 44.97% of them expose or suggest a specific gender/sexual identity or interest likely to be linked to the user,” the study’s abstract stated.

Jusopovic’s finding “IPTC special instructions” simply refers to a special kind of coded watermark that Facebook adds to tag the image in its own coding. These are where the “tracking” comes into the picture – those tags can be read later. Meaning, Facebook will now how the data of people who had downloaded the image, reuploaded it, and all other information that can be linked to the movement of the image across the internet.

What does it imply?

Adding coded watermarks to images is not new tho. It can be used by the company in arbitrating copyrights claims, or providing better user service, and even in targeting the right target market for advertising. However, as Jusopovic states, what he found was a “shocking level of tracking.”

According to one analyst, the metadata has been added since 2016 and “contains an IPTC block with an ‘Original Transmission Reference’ field that contains some kind of text-encoded sequence. This coding method lets Facebook “know it has seen the image before when it gets uploaded again,” explained a user on Reddit. “It is yet another way to learn associations between people. Person 1 uploaded a bunch of the same photos Person 2 uploaded, let’s show them both all the same advertisements!”

But it needs to be clarified that there is no active tracking that it is happening. With “tracking,” Jusopovic means that Facebook and other people with the right tools can access all the metadata in a photo including the crumbles it picks up as it moves from one platform to another. However, this too carries a lot of security of implications.

“Hidden data can be predictably transmitted through social network images with high-fidelity – and AI can hide that data in plain sight, at large-scale, and beyond human visual discernment, making steganalysis and other countermeasures difficult,” Zack Allen from ZeroFOX.

Nonetheless, the discovery of Jusopovic is proof that Facebook is yet to make do with its promises to protect people’s data as well as make sure that people are aware of what happens to their data. This promise follows the high-profile case involving Facebook and Cambridge Analytica, where the social media giant was slapped with a five-billion fine by the FTC.

Continue Reading

Technology

Educational Institutions And Services Are Targets Of Recent Data Breaches

Two separate data breach have affected 62 US colleges and more than 7 million users of the K12.com, with vulnerabilites already exploited by hackers.

Published

on

Photo by Element5 Digital on Unsplash

Education institutions have been victims of two separate data breaches and hacker attacks that have exposed millions of student data to the prying eyes in the wild.

For one of the exposure, more than 62 US colleges have been breached by a hacker that has exploited a vulnerability in an enterprise resource planning (ERP) web app; the U.S. Education Department confirmed this week.

Ellucian Banner Web Tailor, a module of the Ellucian Banner ERP, which allows universities to customize and design their hope-page websites and online applications, has had a vulnerability that was exploited by a hacker who made fake user profiles that are “almost immediately for criminal activity.”

Banner Web Tailor is a web tool, made for higher education institutions, that provides registration, curriculum management, advising, administration, and reporting functionality. Students can access and change their registration, graduation, and financial aid information. It is also used by professors and teachers to input grades which the students can then view online. It is used by hundreds of institutions, many of which have opted to use the Single Sign-on Manager to participate in CAS- and SAML-based single sign-on services.

Joshua Mulliken, a cybersecurity researcher discovered a vulnerability in the authentication mechanism used by the two modules earlier this year. This vulnerability allowed a hacker to hijack students’ web-sessions and take over their accounts. Ellucian fixed the vulnerability in May, and public disclosure was published, by both the researcher and NIST.

“An improper authentication vulnerability (CWE-287) was identified in Banner Web Tailor and Banner Enterprise Identity Services. This vulnerability is produced when SSO Manager is used as the authentication mechanism for Web Tailor, where this could lead to information disclosure and loss of data integrity for the impacted user(s). The vendor has verified the vulnerability and produced a patch that is now available. For more information, see the postings on Ellucian Communities,” reads the public disclosure of the vulnerability.

According to the announcement made by the US Education Department, hackers have already started exploiting the said vulnerability. “The Department has identified 62 colleges or universities that have been affected by the exploitation of this vulnerability,” officials said.

We have also recently received information that indicates criminal elements have been actively scanning the internet looking for institutions to victimize through this vulnerability and developing lists of institutions for targeting with this exploitation.”

The attackers, as said by the officials for the Education Department, “leverage scripts in the admissions or enrollment section of the affected Banner system to create multiple student accounts.” One victim reported that the attackers created thousands of fake accounts over days, with around 600 accounts created within 24 hours.

K12.com exposes users’ data

Meanwhile, a recent data breach involving K12.com has compromised more than seven million students’ data who use one of the company’s programs, leaving the data accessible to anyone online.

In June 25, 2019, Comparitech and security researcher Bob Diachenko uncovered the exposure when they found an unprotected MongoDB out in the open as they scan for unsecured databases.

The exposure affected K12.com’s A+nyWhere Learning System (A+LS), which is used by more than 1,100 school districts. The database has 6,988,504 records containing students’ data. The information held within each file included:

  • Primary personal email address
  • Full name
  • Gender
  • Age
  • Birthdate
  • School name
  • Authentication keys for accessing ALS accounts and presentations
  • Other internal data

“In this instance, an old version of MongoDB (2.6.4) was being utilized. This version of the database hasn’t been supported since October 2016. What’s more, the Remote Desktop Protocol (RDP) was enabled but not secured. As a result, the database was indexed by both the Shodan and BinaryEdge search engines. This means the records contained in the database were visible to the public,” said the researchers.

The researchers were able to reach out to K12.com, and a representative from the company said: “K12 takes data security very seriously. Whenever we are advised of a potential security issue, we investigate the problem immediately and take the appropriate actions to remedy the situation.”

While the danger that comes from the exposed data was not as huge as the first data breach involving the world’s educational system, the researchers said that it carries with it some implications.

“While the leak of this information isn’t as bad as, for example, the exposure of financial data or Social Security numbers, it does have its implications. These pieces of information can be used to target individual students in spear phishing and account takeover fraud. Having their school name made public could potentially put students at risk of physical harm,” they said.

Continue Reading

Technology

Morpheus Chip Is Almost Impossible To Hack Says Researchers

Much like the Greek god, Morpheus chip is a hacker’s nightmare.

Published

on

"Memory Chip - Intel, 1702A-6, 1970s" by Matilda Vaughan is licensed under CC BY 4.0

As cyber criminals up their game in terms of deriving different sophisticated techniques to compromise systems, steal funds, and take over identities, a group of researchers has developed what some media outlets called unhackable chip that is designed to confuse hackers and prevent them from infiltrating systems at the microprocessor level.

Researchers from the University of Michigan call it “Morpheus,” and it said to be “vulnerability tolerant” as it blocks potential attacks by confusing hackers and making it difficult for them to hack at the chip level.

Morpheus is backed by the famous US Defense Advanced Research Projects Agency (DARPA) and is a new chip architecture that bridges the gap between program and machine-level semantics making a system’s firewall impenetrable.

“Attacks often succeed by abusing the gap between program and machine-level semantics– for example, by locating a sensitive pointer, exploiting a bug to overwrite this sensitive data, and hijacking the victim program’s execution,” reads the study’s abstract.

Most common attacks today use malware to trick systems into misusing basic programming possibilities such as permissions and code injection, or into manipulating unusual states.

“In this work, we take secure system design on the offensive by continuously obfuscating information that attackers need, but normal programs do not use, such as [the] representation of code and pointers or the exact location of code and data,” the researchers added.

Just like the Greek god, Morpheus can manipulate program values

According to the developers of the Morpheus, the new chip architecture combines two powerful protections: ensembles of moving target defenses and churn.

The first layer of protection which is ensembling moving target defenses, the chip architecture randomizes key program values (e.g., relocating pointers and encrypting code and pointers) “which forces attackers to probe the system before an attack extensively.”

This means that much like Morpheus, the Greek god of sleep, the chip architecture can make fundamental program values (dreams) to shapeshift to make it harder for future hackers to distinguish them and exploit their vulnerabilities.

The second layer of protection brought by the new chip architecture is the churn, which transparently re-randomizes program values running underneath the system. “With frequent churn, systems quickly become impractically difficult to penetrate,” they said.

Furthermore, Morpheus is also designed to perform both the protection processes, re-randomizing program values every 50 milliseconds, which is faster than any hacker can catch up with, making it highly difficult to locate.

“Each moving target defense in Morpheus uses hardware support to individually offer more randomness at a lower cost than previous techniques. When ensembled with churn, Morpheus defenses offer strong protection against control-flow attacks, with our security testing and performance studies revealing:

  • high-coverage protection for a broad array of control-flow attacks, including protections for advanced attacks and an attack disclosed after the design of Morpheus, and
  • negligible performance impacts (1%) with churn periods up to 50 ms, which our study estimates to be at least 5000x faster than the time necessary to possibly penetrate Morpheus,” they wrote in their study.

One of the proponents of the study explained how the Morpheus chip works using analogies he derived from solving the Rubik’s cube puzzle.

“Imagine trying to solve a Rubik’s Cube that rearranges itself every time you blink. That’s what hackers are up against with Morpheus. It makes the computer an unsolvable puzzle,” University of Michigan’s Todd Austin explains to journalists.

The researchers said that another way of understanding how Morpheus works is that it is a low-level version of a standard protection technique called Address Space Layout Randomisation (ASLR).

Nonetheless, the researchers said that Morpheus chip architecture is not “unhackable.” Other techniques could fight off its protective mechanisms as it does not address all forms of attack. However, the innovation will reduce the attack surface at the very least.

“Looking ahead, we see great potential for EMTD technologies,” the researchers said. “Beyond control-flow attacks, we envision that a similar approach could be adopted to protect against side-channel attacks, timing attacks, Rowhammer attacks, and even cache attacks. To address each of these additional challenges, we will explore what assets the attacker needs and then develop efficient mechanisms to boost uncertainty and stifle attacks,” they added.

Continue Reading

Trending