AI Knowledge Tool Sphere For Wikipedia by Meta
July 29, 2022
Given how much criticism Meta gets for perpetuating misinformation, it is no surprise the company is putting effort into a cutting-edge, open source tool to test the veracity of information online. It is puzzling, though, that the company has no plans to use the tool on Facebook, Instagram, or Messenger, at least not yet. We learn from TechCrunch, “Meta Launches Sphere, an AI Knowledge Tool Based on Open Web Content, Used Initially to Verify Citations on Wikipedia.” The article includes a very brief video illustrating Sphere at work. Reporter Ingrid Lunden also discusses how Wikipedia will use the tool and the relationship between the two companies, so see the write-up for those details. We find this part interesting:
“Meta believes that the ‘white box’ knowledge base that Sphere represents has significantly more data (and by implication more sources to match for verification) than a typical ‘black box’ knowledge sources out there that are based on findings from, for example, proprietary search engines. ‘Because Sphere can access far more public information than today’s standard models, it could provide useful information that they cannot,’ it noted in a blog post.”
Ah, but is bigger really better? Not necessarily. Lunden continues:
“By open sourcing this tool, Meta’s argument is that it’s a more solid foundation for AI training models and other work than any proprietary base. All the same, it concedes the very foundations of the knowledge are potentially shaky, especially in these early days. What if a ‘truth’ is simply not being reported as widely as misinformation is? That’s where Meta wants to focus its future efforts in Sphere. ‘Our next step is to train models to assess the quality of retrieved documents, detect potential contradictions, prioritize more trustworthy sources — and, if no convincing evidence exists, concede that they, like us, can still be stumped,’ it noted.”
That is not ominous at all. Left unmentioned are issues of bias, which tend to be worse the more open a data set is. Will Wikipedia eventually come to regret pairing up with the Zuckbook?
Cynthia Murrell, July 28, 2022
Facebook Sunset: A Rush to Judgment? Nope. Bus Left Already
July 27, 2022
I read a variation on the Chicken Little story with an infusion of Humpty Dumpty. Sound interesting? Just navigate to “Sunset of the Social Network.” Pretty interesting because this is a Silicon Valley type cheerleading outfit realizing that their outfits don’t match those from the ESPN Cheerleading Championships for 2022. How does one quickly fix a fashion faux pas? Easy. Just claim that social networks are even less trendy than the white pants and red sweaters with Mountain View and Palo Alto logos stitched on the polyester.
So what’s the future? If you haven’t figured it out, the answer is TikTok and recommendations to drive memes, advance fun activities like jumping off roofs, and making wlw videos for middle schoolers. Yep, the future. Why not toss end a couple of references to the Decline and Fall of the Roman Empire?
The write up says:
Under the social network model, which piggybacked on the rise of smartphones to mold billions of users’ digital experiences, keeping up with your friends’ posts served as the hub for everything you might aim to do online. Now Facebook wants to shape your online life around the algorithmically-sorted preferences of millions of strangers around the globe.
On the surface, this seems to be what Zuckbook is trying to achieve. Irritating the Kardashians was a knock on effect.
The write up points out that digital dinobabies are a bit clumsy when snow falls:
Rivals tried and failed to beat Facebook at the social network game — most notably Google, with multiple forgotten efforts from Orkut to Google+.
I haven’t forgotten Orkut. That misstep illustrated a genetic flaw in Google’s DNA. Not only could Google not solve death, it couldn’t solve Facebook nor, more recently, Amazon’s gobbling a very large chunk of product search. (Presumably an able Verity alum will redress that issue with information gene splicing. Well, that’s the theory.
Here’s the passage I quite liked:
But the era in which social networking served as most users’ primary experience of the internet is moving behind us. That holds for Twitter, Facebook’s chief surviving Western rival, as well. Twitter never found a reliable business model, which opened it up to an acquisition bid by Elon Musk. Whatever the outcome of the legal fight now underway, Twitter’s future is cloudy at best. … The leadership of Meta and Facebook now views the entire machine of Facebook’s social network as a legacy operation.
Yowza. The very thing that helped make Silicon Valley punditry the next big thing has moved on. Apparently the email has not yet reached Medium and Substack yet. It has, in my opinion, reached Buzzfeed’s senior team and is probably in the in box of a number of other information outlets. That’s just a guess on my part, however.
And what’s the future? The answer is revealed:
All this leaves a vacuum in the middle — the space of forums, ad-hoc group formation and small communities that first drove excitement around internet adoption in the pre-Facebook era. Facebook’s sunsetting of its own social network could open a new space for innovation on this turf, where relative newcomers like Discord are already beginning to thrive.
News flash!
That era has already arrived, and it is morphing, innovating, and invigorating interesting new mechanisms of informationization. Want an example? Okay, CSAM on Telegram. I address this disturbing activity in my luncheon talk at the upcoming Federal Law Enforcement Training Center talk. The downturn in Dark Web activity illustrates a trend building over the last six years.
Facebook and the Silicon Valley real news folks now realize something has changed. Too late? For some, yes.
Stephen E Arnold, July 27, 2022
Hasta La Vista News. Et Tu Facebook?
July 26, 2022
Many Internet users use Facebook as their main news platform. One of the problems of using Facebook is that it builds a confirmation bias, conspiracy theory, and misinformation trap. Facebook has garnered a lot of bad press as a news purveyor throughout the COVID-19 pandemic. Nieman Lab explains that Facebook might be done sharing the news: “Facebook Looks Ready To Divorce The News Industry, And I Doubt Couples Counseling Will Help.”
Facebook could never officially divorce itself from the news, but the social media platform is moving as far away from the news as possible. Facebook pays news outlets millions of dollars each year to post stories in its “News” feature that users can access for free. Facebook had three-year contracts with many news outlets that will soon expire. Facebook is interested in reinvesting the news fees into short video producers like TikTok.
Speaking of TikTok, Facebook has informed its workers to make the platform resemble TikTok. Facebook will now prioritize posts on feeds regardless of their origin, instead of posts users follow. Facebook’s app will continue to feature content from people, then show items recommended by its discovery engine.
The breakup might be mutual:
“So on one hand, Facebook might stop writing checks to news publishers, having found they don’t make its PR problems go away. And on the other, Facebook wants to demote what little news still remains in its primary feed, having found that it doesn’t keep users engaged as much as an algorithm-generated stream of random videos.
This is what a breakup looks like. Facebook was not originally intended to be the world’s largest distributor of human attention to news stories. It became that, circa 2015. But that responsibility became a nuisance, and it’s spent the past seven years walking away from it.”
Maybe this will be the end of conspiracy theorists with a megaphone.
Whitney Grace, July 26, 2022
The Engadget Facebook Entanglement
July 22, 2022
Engadget is a Silicon Valley type of “real” news outfit in my opinion. The online information service published “Fired Employee Claims Facebook Created Secret Tool to Read Users’ Deleted Messages.” The main idea is that Engadget presents information illustrating some fancy dancing at Facebook. The source is a “former employee.”
The write up reports:
a fired Meta employee who claims the company set up a “protocol” to pull up certain users’ deleted posts and hand them over to law enforcement.
Interesting.
I have heard that Facebook, like some other online outfits with oodles of data, has a procedure in place to respond to legally-okayed requests for certain information. I have also heard that Facebook, like other big time information outfits, does not have the resources to respond to requests as quickly as some officials desire. At one conference, I heard a remark that suggested some Facebook professionals were often busy with other prioritized tasks. The legally-okayed requests were placed in a queue. Eventually the Zuckers got to the requests.
Pace, energy, and responsiveness — these are often the hallmarks of a successful investigation. When absent, the momentum is embedded in digital Jell-O. The treat comes in one flavor: Bureaucratic blueberry, a tangy and bitter treat.
The write up points out this allegation presented in a former employee’s complaint:
a Facebook manager briefed Lawson [a former employee who presented the information] on a new tool which, “allowed them to circumvent Facebook’s normal privacy protocols in order to access user-deleted data.
The article explains Facebook’s alleged actions, its software tools, and the former employee’s actions regarding a method for viewing deleted content.k
Several questions:
- The question is, “Are data really deleted or are pointers removed and the data remain in the system?” In my experience, “removing” data can be a tricky and resource intensive process.
- Another question which occurred to me was, “Is this alleged behavior of the Zuckbook surprising based on the firm’s behaviors manifested in Congressional hearings over the last five years? I know that I was not surprised.
- What are the consequences for Facebook if the allegations are in fact true? An engineer can explain that such and such a tool was little more than a modified utility routinely used to determine what content is consuming storage space allocated for a particular data table. Will the legal eagles be able to resolve a repurposed utility designed to investigate storage space?
I am also intrigued with Engadget’s interest in Facebook. My question is, “What’s the entanglement at a distance between these two remarkable companies?” Engadget finds leakers. Leakers find Engadget. Facebook stories are like replays of events on “Live at Five” TV news programs. Repetitive and routine whether “real” or cooked up like a presenter’s recollection of an event like taking fire in a helicopter flying in a war zone.
Stephen E Arnold, July 22, 2022
Meta: Trying Not to Zuck Up
July 20, 2022
Meta is the umbrella company for Facebook and Instagram. The company created the Oversight board to monitor appeals for content moderation on the platforms. The BBC examines the Meta and the banned content in: “Meta Board Hears Over A Million Appeals Over Removed Posts.” The majority of the disputed posts were from Canada, Europe, and the United States. They contained violent, hate speech, or bullying content.
The Oversight Board published twenty cases of appealed content and ruled against Meta in fourteen of them. Some of the cases were: photos of female breasts in a breast cancer post, a photo of a dead child with text about whether it was right to retaliate against China for how it treats Uighur Muslims, and the decision to ban Donald Trump after the January 6 rots. The board overturned banning the breast and dead child images, but supported the Trump decision.
The Oversight Board was originally going to review 130 cases, but Meta agreed that it was wrong removing content on fifty-one of them.
“Board director Thomas Hughes said it looked for “emblematic” cases with “problematic elements” to take on. He added that the categories of hate speech, violence and bullying were “difficult-to-judge issues” – especially for automated systems. ‘Also in many of those cases, context is extremely important,’ he said.”
The Oversight Board released its first annual report covering October 2020-December 2121. Anyone can appeal a decision about removed content. During the first period, 1.1 million cases were received, 2,600 cases are reported a day, and 47 of them came to the board. Most of the complaints came from western countries. Ninety-four percent of the requests were to restore content mostly a user’s posts.
The Oversight Board is compared to a supreme court for Meta and Mark Zuckerberg formed it. Meta pays for its costs, but it operates separately. Its members include human rights activists, lawyers, academics, and journalists. During the appeals session, the board made 86 more recommendations, including translating policies into more languages and being more specific about what constitutes hate speech.
Whitney Grace, July 20, 2022
Meta or Zuckbook: A Look Back to 2020 and 2021 and Years of Human Rights and Other Stuff Progress
July 18, 2022
Meta or the Zuckbook is into human rights. The evidence is a free 83 page report called “Meta Human Rights Report: Insights and Actions 2020-2021.” The document covers in order of presentation:
An Executive Summary (~ seven pages)
Meta’s Human Rights Work in Practice (~ two pages)
Table of Contents with the book beginning on page 13 (yeah, I wondered about the numbering too)
Human Rights Policy Timeline
Part 1: Meta’s Human Rights Commitments (~ 11 pages)
Part 2: Meta’s Human Rights Policy in Practice (28 to 82 or 54 pages)
A Final Note.
The content of the report is interesting. I found a couple of statements which caused me to take up my trusty True Blue color marker. May I share what I circled?
We seek to embed our commitments in a governance model which supports integration of our human rights work with ongoing activities and policies on civil rights and Environmental, Social and Corporate Governance (ESG) efforts, as part of the company’s culture, governance, decision making processes and communication strategies.
Seek and you will find I suppose.
Simply put — we seek to translate human rights guidance into meaningful action, every day.
Yep, another notable “seek.”
In these circumstances we seek to promote international human rights standards by engaging with governments, and by collaborating with other stakeholders and companies.
Okay, seek. How about a quick visit to FreeThesaurus.com for some help?
We also have technical mechanisms in place to mitigate and prevent third parties from accessing data from Meta, through proactive and reactive measures like prevention, deterrence, detection and enforcement.
Do Israeli intelware companies have systems and methods to obviate these super duper data slurpers? “Senator, that you for the question. I will send that information to your office” may be the response to a Congressional questioner.
I enjoyed this quote from the sci fi icon Isaac Asimov:
“The saddest aspect of life right now is that science gathers knowledge faster than society gathers wisdom.”
Here’s my take on Facebook-type social media:
Nothing tears apart a society than ill-managed, ad-centric social media.
I am not Isaac Asimov, but I think I am correct in my observation. Enjoy the “looking back” report from the estimable virtual reality social ad selling heir to MySpace and Friendster. Will Facebook share a similar fate? Gee, I hope not. I am interested in learning if Isaac Asimov’s quote applies to Facebook in 2022:
You don’t need to predict the future. Just choose a future — a good future, a useful future — and make the kind of prediction that will alter human emotions and reactions in such a way that the future you predicted will be brought about. Better to make a good future than predict a bad one.
Did the Meta Zuck thing predict I would sit in my chair with a headset on, interacting with what may or may not be humans, instead of meeting in a coffee shop or an office conference room and talking to a live person? What’s up in 2022? Wow, more Zuckster stuff.
Stephen E Arnold, July xx, 2022
A Digital Medicine Show: Zucking Again?
July 13, 2022
Despite what Meta would have us believe, we learn from MIT Technology Review that “Facebook Is Bombarding Cancer Patients with Ads for Unproven Treatments.” Is Facebook actively prioritizing ad dollars over the wellbeing of its users, or is it just too cheap to hire enough fact checkers to keep the ads from slipping past its algorithm? Either way, research shows targeted ads for ineffective or even harmful treatments flooding cancer patients’ Facebook feeds. These nostrums can be extremely expensive, are usually not covered by insurance (for good reason), and often require travel out of the country. To make matters worse, chasing such bogus cures can steer patients away from effective medical care or, if they are pursuing both, actively interfere with treatments like chemotherapy. Reporter Abby Ohlheiser writes:
“Evidence from Facebook and Instagram users, medical researchers, and its own Ad Library suggests that Meta is rife with ads containing sensational health claims, which the company directly profits from. The misleading ads may remain unchallenged for months and even years. Some of the ads reviewed by MIT Technology Review promoted treatments that have been proved to cause acute physical harm in some cases. Other ads pointed users toward highly expensive treatments with dubious outcomes.”
See the article for details on ad campaigns from the quacks at CHIPSA and Verita Life. We also learn about consequences patients can face if they believe the carefully designed claims from such outfits:
“The danger is not simply that the treatments are unproven or ineffective. Some alternative cancer treatments advertised on the platform can cause physical harm. … Unproven treatments can also interact poorly with conventional treatments like chemotherapy should a patient decide to pursue alternative care on their own. Moreover, simply delaying the start of proven therapies by detouring into unproven ones can allow the cancer to advance, complicating and diminishing the effectiveness of further treatment.”
Facilities make a lot of money from these rackets, and it does not come from the pockets of big insurance. For example, we learn:
“One recent GoFundMe campaign for a cancer patient seeking treatment at CHIPSA included a screenshot of a bill for the ‘base amount’ he’d have to pay. It was $36,500 for three weeks of inpatient care in Mexico. That cost would increase once the facility decided on a treatment plan.”
Ohlheiser graciously states Facebook has done better in the last few years to stem the tide of false health claims, but suggests more fact checkers would help it do better. We agree that would be a good solution, if only the company were motivated to use it.
Cynthia Murrell, July 13, 2022
The Zuck Thing: We Capture Data. You Are Not Permitted to Capture Facebook Data
July 12, 2022
Does anyone remember Eric Schmidt, the alleged Google adult. That senior manager was outraged about open source information about him. You can refresh your memory by scanning this CNet article. Now the Zuckster is annoyed with third parties downloading data available to anyone from Facebook aka Zuckbook. The idea! Take. Info. From. Facebook. “Meta Sues a Site Cloner Who Allegedly Scraped over 350,000 Instagram Profiles” reports:
On Tuesday [July 5, 2022], the company filed separate federal lawsuits against a company called Octopus and an individual named Ekrem Ate?. According to Meta, the former is the US subsidiary of a Chinese multinational tech firm that offers data scraping-for-hire services to individuals and companies.
I find the big company reaction amusing. It is not as hilarious as the Ernsy & Young professionals who allegedly took short cuts to pass an ethics test. Nor is it up to the level of MBAism demonstrated by some McKinsey professionals in their fancy dancing about the “opioid” engagements. But it is pretty funny to me.
It might be a tough call if I were asked to identify the most surveillance oriented big tech company. I have some possible outfits in mind; for example, possible Amazon, Apple, Zuckbook, Google, Microsoft and a few others.
I wonder if Eric Schmidt will give the Zuck some tips on dealing with the outrageous behavior of third parties who gather and recycle data on a public Web site. Imagine, using software to obtain digital information.
The one percent don’t cotton to the “other percent” behaving in a manner offensive to the masters of the datasphere in my opinion.
Stephen E Arnold, July 12, 2022
Meta Plugin Nabs Sensitive Data from Healthcare Websites
July 11, 2022
Never one to let privacy concerns stand in its way, Meta (formerly known as Facebook) has apparently been using its Meta Pixel plugin to collect some of the most delicate data from its healthcare clients. The Markup investigated the matter and shares the results in, “Facebook Is Receiving Sensitive Medical Information from Hospital Websites.” Reporters Todd Feathers, Simon Fondrie-Teitler, Angie Waller, and Surya Mattu tell us:
“A tracking tool installed on many hospitals’ websites has been collecting patients’ sensitive health information—including details about their medical conditions, prescriptions, and doctor’s appointments—and sending it to Facebook. The Markup tested the websites of Newsweek’s top 100 hospitals in America. On 33 of them we found the tracker, called the Meta Pixel, sending Facebook a packet of data whenever a person clicked a button to schedule a doctor’s appointment. The data is connected to an IP address—an identifier that’s like a computer’s mailing address and can generally be linked to a specific individual or household—creating an intimate receipt of the appointment request for Facebook.”
The investigation also found Meta Pixel installed on seven health systems’ patient portals, five of which they documented sending real, password-“protected” patient data straight to Facebook. The detailed article shares some examples of data they caught Pixel collecting and several requisite CYA statements from hospitals and Meta itself. It also explains why hashing is an inadequate, and even duplicitous, privacy measure.
Though Meta-book is not subject to HIPAA regulations, hospitals and healthcare systems certainly are. The authors report:
“Former regulators, health data security experts, and privacy advocates who reviewed The Markup’s findings said the hospitals in question may have violated the federal Health Insurance Portability and Accountability Act (HIPAA). The law prohibits covered entities like hospitals from sharing personally identifiable health information with third parties like Facebook, except when an individual has expressly consented in advance or under certain contracts. Neither the hospitals nor Meta said they had such contracts in place, and The Markup found no evidence that the hospitals or Meta were otherwise obtaining patients’ express consent.”
The authors are unsure how Meta is using this ill-gotten data, observing it could be for advertising, algorithm-training, or other profitable purposes. For its part, the company points to its health information filtering system that is supposed to block such sensitive data from its own grasp. Though, even it admits, that filter is “not yet operating with complete accuracy.” You don’t say. To their credit, several hospitals and health systems removed the plugin once the Markup showed them its findings. So some entities value patient privacy, or at least respect the threat of HIPAA-related prosecution. And it seems other enterprises never will.
Cynthia Murrell, July 11, 2022
Management Theory: Zucking for Efficiency
July 7, 2022
I read “Mark Zuckerberg: We’re Turning Up the Heat at Meta So Employees Will Quit.” I wonder if the blue chip consulting firms — once some of their legal problems subside will embrace Zucking? “Zucking” is, according to the article, the new management mantra at the estimable social media company. Not long ago, imitation became the engine of innovation at the firm: A Discord clone, the TikTok envy resolver, and big boy decision to step back from some high school science club projects. Now, according to the cited article:
Facebook parent Meta wants to cut ties with workers who can’t meet newly raised performance expectations as the company prepares for an economic downturn…
Does anyone remember MySpace? Is this the specter of past social media high fliers accepting the idea of gravity. I am thinking gravity of legal situations, gravity of some advertising customers, and the Jupiter pull of TikTok attracting former Zuckbookers the way our French bulldog does fleas, ticks, and mud. Unstoppable? For sure when French bulldogs are in the woods.
The write up adds another less than positive observation:
Zuckerberg indicated that Meta plans to slow its hiring plans for engineers by at least 30% this year – adding roughly 6,000 or 7,000 workers rather than the 10,000 it initially expected to hire. Some roles that are currently empty will stay unfilled as Meta dials up pressure on current employees.
Wow, a new management concept to galvanize companies finding themselves in a storm: Zucking. When will Zuck Management Practices become a book? Oh, how is that investigation of a former senior manager going? Efficiently I presume.
Stephen E Arnold, July 7, 2022
-
- Subscribe to Beyond Search
Feature archive
News archive
-
