Mediaboss Marketing

AUSTIN, Texas--A former Facebook (now Meta) employee-turned-whistleblower shocked the world when she disclosed tens of thousands of internal documents she took from the company.

Frances Haugen exposed more than20,000 pages of internal documents from the social networking giant.

What is contained in them provided information on many problems,such as Facebooks algorithms to select items to pop up and rank them, which drew huge criticism for accelerating social divides and causing adverse psychological effects on youths.

In a recent interview with The Asahi Shimbun, Haugen spoke about how people should deal with giant social media services. She emphasized that Facebook should fulfill its responsibility and suggested measures for more transparency.

Excerpts of the interview follow:

Question: You pointed out that after Facebook changed its algorithm in 2018, more anger and false information became easier to spread. What surprised me when I read the Facebook documents was that employees had been discussing this for years and couldnt make a decent improvement.

Haugen: There is a document in November 2020, where Facebook says, were going to start valuing anger less. If you look at the comment thread on that document, its huge. People were like, Oh, finally, weve only been saying this for 18 months. And about six weeks later, they released the final weights, where theyre like, OK, this is what were actually going to do for the recalibration, and that had gone away. If you look at the comments on that, everyone is like, what happened? You told us six weeks ago, its going to happen and now youre not going to do it, whats up?

Q: In a congressional hearing last September, you said that Mark Zuckerberg, Facebooks CEO, decided that they would not change the algorithm.

A: I think the issue is this question on responsibility. What I bring up often is the idea that its hard to admit you have power. Because in the process of admitting you have power, you also have to acknowledge that you have responsibility. Its much easier to say our hands are tied, were the defenders of free speech, theres nothing we can do, than to come in and say we actually have all these different levers that dont involve content. They involve things like how easy should it be to reshare? Twitter makes you click on a link before you reshare. Should Facebook do that? What about waiting 30 seconds?

Because those have direct trade-offs and little slivers of profit. I think its a thing where Facebook doesnt want us realizing that thats the conversation we have to have. Because the way the system is today is much more profitable. And not actually even much more, its like, we could probably reduce misinformation by 75 percent. For 1 or 2 percent of profits, the price is not insane. Its not a huge thing.

Q: One of major issues the Facebook documents pointed out is the lack of measures to moderate content in developing countries. Could you elaborate?

A:For example, Ethiopia has 120 million people. They have six major language families and 95 dialects. Its incredibly linguistically diverse. Facebooks current strategy does not work there. Because censorship doesnt scale. With censorship, you have to rebuild those safety systems in every individual language in the most fragile places in the world.

In places like Ethiopia, Libya, Syria, every single time, the top 10 most popular content was horrible. It would be accusing the opposition of mutilating children, it would be severed heads, it would be incredibly sensationalistic.

I think people who live in places that have a rich internet forget that. Id say the majority of languages in the world, 80 to 90 percent of the content available on the internet, in that language is only available on Facebook. And we just forgot this when we speak English or Japanese or whatever.

Q: Why did you decide to come out as a whistleblower?

A:I didnt become convinced. I needed to get information to the public until after they resolved civic integrity. I have an MBA, I went to Harvard. And I literally took a class on change management. Like the study of how you have organizations change, its very well established. You have to have institutional support within your company, where you say, heres a change center. And when Facebook dissolved the civic integrity group, it was a demonstration that they didnt want to have critical mass for change inside the company.

I felt that they had missed the lessons of the 2020 U.S. presidential election. The 2020 election happened, and theyre like, Oh, there wasnt blood in the streets. Therefore, we succeeded. I felt like that was ignoring how big the risks were. And if youll notice that Jan. 6th happened right after that. So clearly, we were not out of the woods yet.

Q: You have been working in the tech industry for a long time. Why did you join Facebook?

A:Id worked at three social networks before Facebook. I founded the search team at Google Plus, I worked at Yelp, I helped build their machine learning team. And I was the lead product manager for ranking at Pinterest. Im not trying to brag here. But theres not a lot of people in the industry that are like both algorithmic specialists, who also look at the human impacts of the system. I would guess theres something on the order of maybe 200 or 300 people in the whole industry, that have the depth of experience with these recommender systems that I do, across these kinds of development.

The question I had to answer was, if my fears came true, so I genuinely believe theres like tens of millions of lives on the line in the next 20 years.

Q: We heard that one of the reasons you joined Facebook was your friends experience where he believed in conspiracy theories. Could you tell us about it?

A:Part of why I took the job at Facebook was I think people who have not personally experienced someone getting radicalized, dont understand it. You can really trivialize it. You can say, Well, smart people dont have that happen to them, or educated people dont have that happen to them. Like that happens to other people.

My friend was a friend of my little brother, and thats how I met him. Hes the same age younger than my brother that my brother is younger than me. My mother always wanted a third kid and so I always thought of this guy as like my little brother. I can walk today, Im healthy, because he saved me. Watching him fall into this darkness made me feel so powerless.

Because Im a child to scientists. Theyre academics. I believe that truth is the thing that we can see. If you interact with someone who is falling victim to the echo chamber of misinformation, you realize that humans are so susceptible to the social context of facts. If you are in an echo chamber where the same messages get reinforced, and the actual consensus reality is eroded, you lose the chance to even connect.

These systems are not neutral. As someone falls down the rabbit hole, its not like its pleasant for them either. If you believe your government is trying to poison you, if you believe teachers are trying to hurt kids, like all these things, its not like it makes your day-to-day existence more pleasant. You watch someone fall away from our community, like our clean, our shared reality. If you feel like you cant pull them back, its a real horror. I think theres lots of families that have experienced this, like youve had a relative go into the dark corners of the internet. I think if we treated with more respect for the people who had those experiences, and understood that the platforms are far more responsible for these experiences than the people who fall down these rabbit holes, because the products are designed to be addictive.

Q: What is the best way for the society to treat the algorithms that have such strong power to decide what we are seeing on the Internet?

A:Right now, theres no route for academics for governments to say, Hey, you need to do the following ongoing reporting. If you were a bank, we have other systems that are really opaque and complicated, that drive our lives, that we treat differently than we currently treat social media companies. So, for example, medicine is very complicated, and intimately impacts our lives. We have codes of ethics for doctors because of that. Because we know that they have more information than we do. They have a duty to take care of us. Lawyers, the legal system is complicated. Theres a duty of care there, where they have to look out for us with the things that we know these systems are complicated, and very few people understand.

Facebook is opaque. We each only get a little tiny peephole that is our own individual experience. We dont get to see the aggregate, like tableau. Up until now, Facebook is taking advantage of this, because no one else could do the research. No one else could see anything more than their own little peephole.

Right now, Facebook doesnt have the most viewed content report for every country in the world. They only have it for the United States and they only show you 20 pieces of content, which is crazy. I think the reason they do that is if they show the top 5,000 pieces of content, we would be outraged. The reason they dont show us is they dont want us to see that. The fact that Facebook wont disclose, even when they were asked, shows you how messed up the system was.

Basically, the only question to discuss here is censorship. We are all used to free speech. So were not going to go further than this. In reality, the entire time they had the solutions to keep us safe, that didnt involve censorship. They chose not to do them. They chose to not even let us talk about them.

We have to figure out a different feedback loop, where maybe Facebook has to disclose what they know. Maybe its a thing where theres an aggregate data that could be released. That would give a different incentive for them to actually accept these changes, if they had to publish their misinformation numbers all the time or show their most viewed content.

Q: When you spoke at congressional hearings in the United States and Europe, you called for stronger regulations. What do you think is the most important point?

A:We need to write laws that require Facebook to disclose more information. Because right now, we dont have the public muscle of accountability. Lets say were talking about an oil company. Every year, we graduate worldwide tens of thousands of environmental science majors. These are people who have been trained in like, Heres the ways that we monitor companies to make sure theyre healthy.

Right now, we dont graduate anyone who has a depth of knowledge in the systems that we have, because you cant take a college class on it. You cant get a graduate degree on it. You have to go to these companies and work there to learn about it. That is profoundly dangerous. We need to invest in how we build that public muscle of accountability.

Q: There has been a huge knowledge gap between tech companies and the general public. How can we fill the gap?

A:When you were in high school, you probably took a chemistry class. The thing thats cool about taking a chemistry class is it lets you simulate what its like to be a chemist. You blow up stuff, you breathe in stuff you shouldnt breathe in. You kind of get a sense of what its like to be a chemist. But if you want to be a data scientist, or you want to be an algorithmic engineer, theres no equivalent lab to begin to learn those meta skills and allow you to do those things.

I really want to build that lab bench for social media, where you have a simulated world, where we can come in there and have a conversation about things like when we reshare content, once it gets beyond friends of friends, what happens? If we had a copy and paste, what would happen? Who would win? Who would lose? What would happen? Because on a lot of these things, the questions arent the number bigger or lower. Its about there are some winners and losers. How do we weigh all those trade-offs? Right now, Facebook is resolving those trade-offs in a really simple way. Did profits go up? Did profits go down? Did growth go up? Did growth go down?

If we had 19-year-olds arguing about some of these things, we teach that class at a bunch of different levels, we can compute the numbers for people who arent very mathematical. They could just look at graphs. We could allow someone who wants to be a data scientist actually try to do it themselves. They can build those skills. We need to start thinking about what is that method of education. Because were entering a world where we have to think beyond censorship.

Q: In Japan, its been said that its really hard for whistleblowers, especially women, to come out to the public. Do you have any message for them?

A:I dont know enough things about gender issues in Japan to feel responsible for giving commentary. Thats all I can say. I feel grateful for how much support Ive gotten. I think the advice I give to all whistleblowers is you need at least one person you trust so that you can talk to. Because its true for all whistleblowers. Because fundamentally, the art of being a whistleblower is first you have to believe what youre seeing is real. And thats hard. Because usually those systems are aligned to tell you its not real. Because they can only continue to exist if their employees continue to go along with whatever the frame of reference is.

Q: I know you are optimistic about the future of social media. Why could you be so optimistic?

A:I think why Im optimistic is we know ways to do things in more safe ways. We also can look to the past. Like we love social media, like social media used to be about our family and friends. When you ask people what is Facebook about, people still say that even though you have very little content from your family and friends today. So I think its interesting.

Im optimistic because I see things like decentralized platforms. I see other companies coming along and saying, Lets do this, but do it the right way. I believe that if we just change the incentives on Facebook. Facebook already knows 20 things they could do that would make these platforms safer overnight. So, I believe that we can hold them accountable. Once the incentives change, the behavior will change.

(Marie Louise Leone contributed to this article.)

Excerpt from:
INTERVIEW/ Facebook whistleblower urges legislation for transparency | The Asahi Shimbun: Breaking News, Japan News and Analysis -

But weve been telling the same story about Arpanet and the web for 25 years, and it isnt satisfying anymore. It doesnt help us understand the social internet we have now: It doesnt explain the emergence of commercial social media, it cant solve the problems of platformization, and it wont help us to imagine what comes after.

Todays social media ecosystem functions more like the modem world of the late 1980s and early 1990s than like the open social web of the early 21st century. It is an archipelago of proprietary platforms, imperfectly connected at their borders. Any gateways that do exist are subject to change at a moments notice. Worse, users have little recourse, the platforms shirk accountability, and states are hesitant to intervene.

Before the widespread adoption of internet email, people complained about having to print up business cards with half a dozen different addresses: inscrutable sequences of letters, numbers, and symbols representing them on CompuServe, GEnie, AOL, Delphi, MCI Mail, and so on. Today, we find ourselves in the same situation. From nail salons to cereal boxes, the visual environment is littered with the logos of incompatible social media brands. Facebook, Google, Twitter, and Instagram are the new walled gardens, throwbacks to the late 1980s.

In recent years, it has become commonplace to blame social media for all our problems. There are good reasons for this. After decades of techno-optimism, a reckoning came due. But I am troubled by how often peoplenot platformsare the object of this criticism. Were told that social media is making us vapid, stupid, intolerant, and depressed, that we should be ashamed to take pleasure from social media, that we are hardwired to act against our own best interest. Our basic desire to connect is pathologized, as if we should take the blame for our own subjugation. I call shenanigans.

People arent the problem. The problem is the platforms. By looking at the history of the modem world, we can begin to extricate the technologies of sociality from what weve come to call social media. Underlying many of the problems we associate with social media are failures of creativity and care. Ironically, for an industry that prides itself on innovation, platform providers have failed to develop business models and operational structures that can sustain healthy human communities.

Silicon Valley did not invent social media. Everyday people made the internet social. Time and again, users adapted networked computers for communication between people. In the 1970s, the Arpanet enabled remote access to expensive computers, but users made email its killer app. In the 1980s, the Source and CompuServe offered troves of news and financial data, but users spent all their time talking to one another on forums and in chat rooms. And in the 1990s, the web was designed for publishing documents, but users created conversational guest books and message boards. The desire to connect with one another is fundamental. We should not apologize for the pleasures of being online together.

Commercial social media platforms are of a more recent origin. Major services like Facebook formed around 2005, more than a quarter-century after the first BBSs came online. Their business was the enclosure of the social web, the extraction of personal data, and the promise of personalized advertising. Through clever interface design and the strategic application of venture capital, platform providers succeeded in expanding access to the online world. Today, more people can get online and find one another than was ever possible in the days of AOL or FidoNet.

More:
The Internet Origin Story You Know Is Wrong - WIRED

Instagram again went down for a number of users on Monday, with many rushing over to Twitter to complain and ask what was wrong with the photo-sharing app this time. It has suffered a number of outages recently. Many also wondered if it is time to log off permanently from the app.

Late on Monday, many users on the app were confused when their explore tab suddenly started showing posts that are not usually seen on their feed; most of them were nature-related posts. Others complained they could not send messages on the app and Instagram Stories wouldnt load.

So, this is pretty but can I have my Instagram feed back to normal. I love nature and aesthetics, but give me my memes back, wrote an user sharing photo of her feed, complaining on Twitter. Who called it Instagram glitch and not Mark Zuckerbug, another tweeted, taking a jibe at the companys CEO.

According to Down Detector, the site recorded nearly 1,200 reports in India with users unable to load Reels or failing to upload anything on Stories. The graph showed the outage peaking around 11 pm and things resolving in an hour. Many also wondered if it was time to say goodbye to the app.

However, that didnt stop netizens from having fun and sharing hilarious memes to react to the outage almost every now and then.

Lots of social media apps like Instagram, Facebook, and Twitter have been reporting outage in the recent past quite frequently, leaving netizens unhappy. Moving beyond social networking apps, food delivery services apps like Zomato and Swiggy too faced outage issues last month.

See the original post here:
#InstagramDown: Users vent with memes, wonder if they should uninstall - The Indian Express

Design and sampling

This randomized controlled trial study with parallel groups was conducted on pregnant women referred to Urmia health centers in 20182019. The study was last 12months for execution, from April 2018 to May 2019. This trial was registered in Iranian registry of clinical trials (IRCT) in 05/06/2020 with IRCT20151004024340N15 identifying number. Also this study was verified by the Urmia ethics committee with IR.UMSU.REC.1397.162 code.

The sample size was calculated according to the Michele Bisson [29] study, because of the similarity in the targeted primary and secondary outcomes, where the moderate and vigorous physical activity (MVPA) index was 11.79.5min/day in their control group and 25.420.4min/day in their treatment group after the intervention. With the 90% of study power (Z1=1.28), and the two-sided alpha level at 0.05 (Z1/2=1.96), and a 10% chance of dropout, the sample size was determined 45 subjects for each group. Therefore, ninety subjects were included in the study. The flow diagram of participants has shown in Fig.1.

Flow diagram of entering the study and completing of trial by participants

Inclusion criteria in the study were: the tendency to participate in the study, having an ability for reading and writing, be under 20weeks of gestational age, lack of chronic diseases (i.e., diabetes, hypertension, kidney disease), lack of any medical limitation for improving physical activity by mother, no mental illness and no history of hospitalization (based on the participant's self-report), lack of cerclage and prenatal care, having a smartphone, access to the internet, and residence in Urmia. The exclusion criteria were: having any problems or special diseases in the current pregnancy such as preeclampsia, diabetes, anemia, high-risk pregnancy and cerclage surgery, frequent bleeding during the intervention, premature contractions, reduction of fetal movements, amniotic fluid leakage, diet for a specific disease, and migration from Urmia.

We used stratified sampling method for including subjects from different socioeconomic strata of Urmia. There were three levels of health centers; 1, 2, and 3 according to their social and economic status. Then, two centers from each category were chosen randomly by lottery (drawing). Also participants for each group were randomly assigned with one to one allocation ratio with the block randomization method. First, the blocks were created with the combination of AAABBB by computer and all the possible statuses were identified and an exclusive code was assigned for each. Next, considering the sample size (N=90) and block sizes (S=6), 15 blocks were chosen by a simple randomization method by the researcher. Selected subjects were allocated to control or intervention group consecutively by the study manager. All these stages were done under the provision of a consultant epidemiologist and using Random Allocation Software version 1.0.0.

In this study, demographic, Pregnancy Physical Activity Questionnaire (PPAQ) questionnaires was used and completed. Height measured with a Stadiometer and weight by a Seca scale. Demographic information questionnaire had questions such as age, weight, height, education level, employment status, economic status, BMI, pregnancy age, number of fetuses, number of pregnancies, history of the underlying disease, history of infertility, cerclage, ectopic pregnancy, and specific disease in the current pregnancy.

Standard PPAQ, which is related to physical activity in pregnancy, includes two parts: Part I consisted of information about individual characteristics and Part II includes 32 questions on physical activities. The questionnaire of pregnancy physical activity consists of 4 groups of questions related to activities: at home (16 questions), community (3 questions), and activity in the workplace (5 questions), and sports and entertainment (8 questions). Within the manual of PPAQ the categorization and classification of daily physical activities based on the intensity and duration of them has been explained well. Based on that manual sum of the values of multiplied intensity and duration for some activities within the questionnaire can predict the amount of MET/Min for each category of activities as sedentary, light, moderate and sever that has pointed in detail in the data analysis section.

After selection of subjects in the first meeting, the researcher introduced herself to mothers and explained the study purpose. Informed written consent was then obtained. Later the questionnaires related to personal information and physical activity were filled in. Body mass index (BMI) was calculated and assessed. The mothers weight was checked with minimum clothing and without shoes, using a digital scale with a precision of 100g. Peoples height was measured using a wall gauge with an accuracy of 1cm while standing barefoot by the wall, as the shoulders were in a normal position. BMI was calculated by dividing weight by height squared.

Based on weight status (i.e., normal, overweight, and obese) and through the permuted block randomization, all the pregnant women were randomly assigned to the control and treatment group by Random Allocation Software version 1.0.0. Both groups received individual diets according to BMI, by a nutritionist, who also provided the necessary explanations. All educational content in the format of text, audio, and video files was delivered and provided through of what's app as social media platform. All intervention and control subjects received their specific educational contents according to the prepared study protocol for each group.

The control group subjects were quiet match and nearly similar in properties with the intervention one. They received a routine pregnancy care that was provided by health care centers and were followed alongside with the intervention group in the same time period. Subjects joined a virtual group in a social network (What's app) that was accessible through both mobile, laptop or PC devices for all subjects from each centers in control group. They received prepared individually diets and materials about that how to track its effect on their weight gain during pregnancy. In parallel, the treatment group also joined a virtual training group in a social network (What's app), that was similarly accessible through mobile, laptop or PC devices for all subjects from each centers in intervention group. Just the difference that was for this group, they received special and prepared educational content as written, audio or video materials to improve their physical activity along with an individually designed diet during 16 sessions in 8weeks. The content delivered to the virtual group twice a week and in a specified time period. Each session was about 90min that all educational programs were delivered by an educated and specialized midwife. During different session subjects regarding the advantages of executing appropriate exercises and physical activities during pregnancy, introducing suitable exercises for pregnant women and the manner of their execution was educated. Also in some sessions topics about the necessary precautions and probably health risks with some activities was discussed. Reminder massages was sent twice a week regarding the importance of doing proposed exercises and activities truly, the importance of good adhering to the diet, and announcement about the time of next session and its content. All womens questions were answered both within the group and in private chat based on the type of question. The respondent rate for both groups was %87.8 in average. Mothers weights were recorded after a face-to-face meeting in the fourth week. All delivered massages deposited in virtual group by the end of intervention. Our primary outcome was measuring the amount of changes in the mean daily total physical activity level, while secondary outcomes were measuring the level of changes in other subgroups of daily physical activity alongside with the level of weight gain during pregnancy.

At the end of the intervention, pregnant mothers in both groups filled the PPAQ questionnaires for the second time. The data acquired in this phase were compared with the first one.

Quantitative data were reported as meanstandard deviation and qualitative data were expressed in the form of frequencies and percentages. Activity intensity was measured based on the amount of Metabolic Equivalent of Task (MET). To calculate the total MET/Min of activities, the MET level regarding the intensity of activities was multiplied by the amount of time spent per day. The level of daily activity determined based on the amount of MET/Min obtained for activities and their average per day. The categorization for the level of physical activity was done according to the PPAQ specific manual. An activity with MET/Min less than 1.5 was considered as inactivity or sedentary, with MET/Min 1.53 as light activity, with MET/Min 36 as moderate activity, and with MET/Min more than 6 as a sever activity. An independent t-test was used to compare the data about the groups for comparing them at start and at the end of intervention. A paired t-test was used for comparing changes within groups and comparing before and after data. Also, the chi-square test was used to compare data regard the qualitative variables between the two groups at start of the study. To control the possible influence of confounding factors and better determination of treatment effect, multivariate analysis was done. The significance level for all statistical tests was considered less than 0.05. We used per protocol analysis for comparing the extracted results between study groups. All analyses were performed using SPSS-21 software.

Continued here:
Examination of influence of social media education through mobile phones on the change in physical activity and sedentary behavior in pregnant women:...

A new study has shown a growing trend in which hackers are impersonating popular social networking sites in order to obtain users personal information or payment credentials via phishing attacks.

Check Point Research, the threat intelligence arm of Check Point Software Technologies and a cybersecurity solutions provider, highlighted the brands that were most frequently imitated by cybercriminals in its Brand Phishing Report for the first quarter of this year.

LinkedIn shot to the top for the first time, accounting for 52% of all phishing attempts globally in January, February and March.

This is a significant 44% jump from the previous quarter, when LinkedIn was ranked sixth and was associated with only 8% of phishing attempts.

The other most-imitated companies are DHL (14%), which ranked top in the last quarter, Google (7%), Microsoft and FedEx (both at 6% each), and WhatsApp (4%).WhatsApp remains in the top 10, the research says, and that the messaging application is responsible for nearly one in every twenty phishing-related attacks worldwide.

If there was any doubt that social media would become one of the most heavily targeted sectors by criminal groups, the first quarter has laid those doubts to rest, Check Point Research said, as social media has surpassed transportation companies and technology behemoths such as Google, Microsoft, and Apple as the most targeted category.

The firm stated that its findings are based on data obtained from Check Point ThreatCloud and its Cyber Threat Map, which analyses cyberattacks in real-time around the world.

These phishing attempts are attacks of opportunity, plain and simple.

Criminal groups orchestrate these phishing attempts on a grand scale, with a view to getting as many people to part with their personal data as possible, said Omer Dembinsky, data research group manager at Check Point Software.

Typically, a hacker would attempt to steal user data by sending malicious emails using the brands name as the sender address, taking care to include elements such as the brands logo or colour scheme to make them appear authentic.

The report highlights an example where LinkedIn users were contacted via an official-looking email in an attempt to lure them to click on a malicious link.

Once there, users would be prompted to log in via a fake portal where their credentials would be harvested, the company said, adding that criminals will attempt to imitate the official website by using a similar domain and design.

In another example, a hacker used Maersks branding and a spoofed email address to make it appear as if it was sent from Maersk Notification.

The user was asked to download an Excel document that contained the Trojan known as Agent Tesla, a type of malware that is capable of granting access to a victims computer through a backdoor.

Some attacks will attempt to gain leverage over individuals or steal their information, such as those were seeing with LinkedIn.

Others will be attempts to deploy malware on company networks, such as the fake emails containing spoof carrier documents that were seeing with the likes of Maersk, Dembinsky said.

The brands of shipping companies like DHL and Maersk, he said, are being abused by threat actors to target both companies and consumers due to the rising popularity of ecommerce, and its critical that staff be taught to be more alert when it comes to recognising phishing schemes.

Employees in particular should be trained to spot suspicious anomalies such as misspelt domains, typos, incorrect dates and other details that can expose a malicious email or text message, he said, urging LinkedIn users to be extra vigilant over the course of the next few months.

Read more:
LinkedIn was the top targeted social media network for phishing attacks in Q1 - The Star Online