Google Fires Second AI Ethicist, Changes Diversity and Research Policies
We may earn a small commission when you click or purchase an item using a link on this website.
A recent internal investigation at Google has resulted in the second firing of a founding member of Google’s ethical AI development team. Margaret Mitchell was relieved of her position as co-leader of the ethical AI division for “[violating] the company’s code of conduct and security policies,” Google said in a statement.
Her firing comes just two short months after her colleague and co-leader of the ethical AI team Timnit Gebru was abruptly let go after she attempted to publish research about the unintended consequences of large text processing systems, including those used by the Google search engine. Google did not approve the publication, citing that the paper “ignored too much relevant research.” Gebru had asked for further discussion before retracting the paper, and said that she planned to resign if the company failed to address her concerns. Google refused to meet the conditions and promptly relieved her of her position.
Gebru is a prominent data scientist who specializes in diversity and bias research in AI technology. She is best known for her research that found significant racial bias in facial recognition technology, showing that the algorithms recognize white faces better than black and brown faces. The dispute that led to Gebru’s firing prompted the resignation of two of her coworkers at Google, and sparked a media frenzy that has recently been amplified by Mitchell’s termination.
The investigation against Mitchell found that she had violated company policies by running automated scripts through Gebru’s email while searching for evidence that Google had discriminated against her. Many who witnessed the situation have publicly stated that Gebru was “wronged,” and have suggested that her undue exit from the company may have had a discriminatory basis.
Mitchell tweeted to her followers about her own firing: “I tried to raise concerns about race [and] gender inequity, and speak up about Google’s problematic firing of Dr. Gebru.”
Alphabet’s Google has come under major scrutiny in recent months following the DOJ’s lawsuit accusing the corporation of violating antitrust laws last October. Amid a slew of related lawsuits that paint Google as a monopolistic enterprise, the company has also been the center of a growing debate about its diversity policies and treatment of employees in minority groups. In response to the incidents involving Timnit Gebru and Margaret Mitchell, Google has recently implemented diversity and research policy changes in the hopes of preventing similar situations in the future.
Asus Confirms Launch Date for Zenfone 8, With ‘Mini’ Version Rumored
Asus have announced a product launch that is set for May 12th in which the Zenfone 8 will receive its official unveiling. Rumors have swirled recently that the company plans to also release a mini version of the Zenfone 8 which has been all but confirmed by the launch event’s invite.
As part of the event, Asus released a new splash page on their website featuring the tagline “Big in performance, compact in size”, hinting that there may be a smaller model of the Zenfone 8 in the works.
In addition to the new splash page on Asus’s website, the company also sent physical invites to multiple news outlets. The folks over at PhoneArena reported on their physical invite noting that it came in the form of a screen protector in the dimensions of 5.83 inches x 2.72 inches.
The screen protector has been interpreted as a clue to what the actual dimensions of the Zenfone 8 or potential Zenfone 8 mini would be, and as PhoneArena noted, if the screen protector invite dimensions are the actual size of the phone, it will be just slightly than the iPhone 12 mini, which comes in at 5.18 inches x 2.53 inches.
The Android phone market has seen more compact handsets go out of style. While you still can find smaller Android phones on the market, they typically are lower quality phones or models of larger phones, making it reassuring that Asus promises the Zenfone 8 will be built with some of the best hardware on the market currently. Earlier this month, a Geekbench listing appeared online for an unnamed Asus phone that had the new Qualcomm Snapdragon 888 chip. If this was indeed a tech preview of the Zenfone 8 as many have been speculating, this means the phone should compete spec wise with the Samsung Galaxy S21 Ultra and the OnePlus 9 Pro.
While Asus has existed as a giant for quite some time in the computing market, their entry into the Android smartphone market has been much more quiet. With the Zenfone 8, Asus is looking to capture a sizable niche in the smartphone market: top tier, smaller phones. For now, we will have to wait until May 12th to see the full list of specs from Asus directly for the Zenfone 8, but the current hints towards size, memory and processor all point in the direction that Asus is aiming to compete with the highest end phones currently available for Android.
Google Introduces ‘Android Ready SE’ to Promote the Adoption of Digital Keys, Mobile ID’s, and Digital Wallets with Added Security
While the Android system has garnered wide praise by the Android community over the years for its open nature, the open nature of the system is subject to safety and privacy concerns. Google, however, is always looking at ways to innovate and improve the Android system by making the system faster, more convenient, and most importantly, more secure for their users.
With the expansion of the Android system past just smartphones and into the world of smart devices such as watches, TV’s and even cars, Google has set up the ‘Android Ready SE Alliance’ to offer new Android hardware with improved security functions. The Android Ready SE Alliance builds upon the already existing Secure Element (SE) technology that was first released in 2018 for the Pixel 3 with the Titan M chip and the tamper-resistant key storage for Android Apps through StrongBox.
In the Google Security Blog posting announcing the Android Ready SE Alliance, Google notes that StrongBox and tamper-resistant hardware has become an increasingly important requirement for new Android features related to digital keys for cars, homes and offices, mobile drivers’ licenses, ID’s, ePassports and digital wallet solutions.
While StrongBox and tamper-resistant hardware has become a major need in today’s world as smart devices expand and become a part of daily life, Google believes the SE technology is the perfect security solution for smart devices, as well as identity certification and digital wallet solutions. Google notes that all of these features will need to run on tamper-resistant hardware in order to protect the safety of their user’s data, keys, wallets, and more. The majority of modern Android phones now include the tamper-resistance hardware known as Secure Element (SE) and the company believes that the SE technology is the best path to introduce new consumers to better safety functions.
The aim of the Android Ready SE Alliance is to speed up the adoption of such technologies on Google’s Android platform. Google is allowing developers to develop open-source and ready-to-use applets for SE-capable microchips.
In addition, Google also announced that StrongBox will be applicable to WearOS, Android Auto Embedded and Android TV, and not limited to just Android smartphones or tablets but rather a wide range of Android devices in the Internet of Things device category.
Google has claimed that they currently in development on a few applets for Android devices that will be released according to priority, with mobile drivers’ licenses, identity credentials and digital keys being high in the priority list. With the company announcing that an applet for the Android Ready SE Alliance is already released to be used by their OEM (original equipment manufacturer) partners to produce secure devices, we should begin to see Android Smart TVs and wearables featuring the secure technology in the coming months.
Overall, the aim of the Android Ready SE Alliance is to provide a better guarantee of safety and security to Android users when it comes to using digital wallets, ID tokens, and even keys, allowing for users to do their everyday tasks without ever having to worry about security again.
New York Times Investigation Into Capitol Rioter Data Reveals False Claims of User Anonymity With MAID Technology
An ongoing New York Times Opinion investigation probing the true nature of location tracking via smartphones has led the Times Opinion journalists to a disturbing discovery regarding the privacy of smartphone users in America:
Anonymity does not exist.
The Privacy Project is a series of articles chronicling a growing body of evidence about just how easy it is for mobile and online users to be tracked by private entities. A subset of articles specifically pertaining to cell phone data and the high potential for its misuse tells the story of a data file obtained by the Privacy Project in 2019 containing more than 50 billion location pings from some 12 million Americans across the country. Charlie Warzel, a key contributor to the Privacy Project series, uses the data to demonstrate how few connections must be made in order for a person’s entire identity, along with their day-to-day activities, to be uncovered.
In the first article in the subset, titled ‘One Nation, Tracked’, Warzel offers a comprehensive look at how one dataset quickly revealed enough information to identify the owner of each phone. Those whose identities were discovered from the data include celebrities and government officials with high security clearances in the mix with average Americans. Warzel even released a spinoff article the same day showing a detailed path of a smartphone connected to someone in Trump’s entourage — an alarming example of how easily parties interested in threatening America’s national security can access sensitive information.
Another recent installment of the One Nation, Tracked subseries unveiled a new dataset obtained following the Capitol siege on January 6th that tracked 130 phones pinged inside the building at the time of the riot. The data revealed that about 40 percent of the pinged phones were near the rally stage before making their way toward the Capitol that day, an undeniable link to those who participated in the march that ended in criminal behavior.
“While there were no names or phone numbers in the data, we were once again able to connect dozens of devices to their owners, tying anonymous locations back to names, home addresses, social networks and phone numbers of people in attendance” reports the Times Opinion article regarding the dataset.
The source who provided the data wanted those who committed acts of treason that day to be brought to justice. They also expressed concern for the vulnerability of the data, citing that they could have just as easily used it for their own insidious purposes rather than bringing it to the Times Opinion reporters.
Using this new dataset, analysts at the Times also discovered a remarkable new piece of information included with the pings: a unique ID tied to each smartphone, and implicitly, each user.
This ID, it turns out, is essentially a cookie for mobile devices known as a mobile advertising ID, or MAID for short. MAIDs are intended to allow advertisers to more effectively target potential customers by identifying a user with an “anonymous” ID, and allowing marketers to reach mobile audiences with personalized ads based on their consumer habits.
MAIDs are not exactly a new development; Google introduced its Android ID system in 2013, with Apple not far behind. What the Times demonstrates through its data analysis is that this supposedly anonymous technology makes it even easier for users’ identities to be discovered, and calls into question the integrity of the systems that have become a staple in the everyday lives of billions globally.
All of this information is owned by private organizations in a largely unregulated industry. The Times Opinion article considers how the monetization of user data that has proven to be very identifiable, not to mention alarmingly accessible, could threaten the civil liberties that Americans hold so dear. The Privacy Project highlights growing concerns about surveillance and online privacy laws in America, and hints at an upcoming battle between private data collection entities and regulatory agencies.