YouTube Tries Out Letting Viewers Fix Captions

YouTube is testing a new idea to make the captions on its videos better. They're asking people who watch videos to help by fixing any mistakes in the captions. This help from viewers is being tested on a small number of videos in English.

Here's how it works: When you're watching a video that's part of this test, you can click on the settings icon and choose "subtitles." Then, you can see the video's captions. If you spot a mistake, you click on a pencil icon and type what you think is the correct caption.

Other people watching the video can also help. They can see the corrections you've suggested. If they agree with your correction, they can vote for it by clicking on it.

Right now, these changes won't automatically update the video's captions. But, the person who made the video can see all the suggested fixes and decide if they want to use them. This could make captions more accurate.

Better captions are important for several reasons. They make videos easier for more people to understand, especially if they have trouble hearing. Accurate captions can also help videos show up in search results, since Google uses these captions to understand what the video is about.

For now, YouTube is only trying this out with a few videos in English to see how it goes.

Image: DIW-Aigen

Read next: X Reveals User Engagement Data: Conflicting Numbers Spark Questions
by Mahrukh Shahid via Digital Information World

Viral Social Media Content Sparks Outrage and Threat Perception, Research Finds

According to a study published in Journal of Personality and Social Psychology, it was found out that a sense of being viral produces outrage and threat to the person. When new trends emerge on social media and become viral, they become a source of harmful issues and can range from economic to climate conditions. On social media, virality means the number of likes and comments a post has acquired. When the likes and comments on a post are higher, it creates a sense of panic and concerns and users pose it as a threat.

Any content that is related to threats in society becomes a source of outrage when it becomes viral. These outrages become more powerful if the said threatful content is opposing the views of a society or talking about a political opposition. In the study, the researchers analyzed different posts about climate change, immigration and Covid-19. The researchers found out that the posts that were stating some different point of views were shared widely and attracted more replies that were expressing anger.

The team of researchers did some experiments with 1500 people. The reason for the experiment was to find out how these participants reacted to tweets that were shared a lot and the ones that weren’t shared that much, and to find out how these people would react to tweets according to their own perceptions. In one part of the experiment, the participants were shown tweets that were talking about some political parties that the participants were supporters or not supporters of. The reactions of the participants were observed according to how they perceived danger in the tweets and which responses made them angry. The next part of the experiment was to show these participants some prejudiced tweets and analyze the emotional and behavioral responses of the participants in terms of virality of the tweets.

The researchers recorded the difference of reactions of participants on annoying tweets and tweets that were harmful and prejudiced. They even made a new hashtag to see how the participants will perceive a new trend if it gets viral. The experiment was concluded after assessing which factors angered the participants, as well as which factors made them worry and sad. Participants were also evaluated on the basis of their reactions on viral tweets on the basis of likes and shares and viral tweets on the basis of the person who has tweeted.

From this experiment, the researchers found out that the posts that become viral on social media are more likely to get opposed and produce harmful emotions like anger. The posts who talk about genuine concerns also elicit deep reactions. The researchers say that they will also do an experiment in the future that will be about virality of a post in terms of trending topics and volume of posts about an issue.

Image: DIW-Aigen

Read next: Meta, TikTok, And Snap Pledge To Work Toward Election Integrity As US Presidential Race Draws Near
by Arooj Ahmed via Digital Information World

Meta, TikTok, And Snap Pledge To Work Toward Election Integrity As US Presidential Race Draws Near

It’s election season and as the US braces for another Presidential race, tech giants are rolling up their sleeves to make sure they can handle the pressure.

Meta, TikTok, and Snap were seen uniting with a host of other leading tech firms in regards to the signing of the IFES’s latest guidelines for election integrity.

The goal is to ensure all companies are working toward making sure there’s limited misinformation spread and greater trust in data content being published online. Only then can it give rise to a healthier ecosystem.

This new agreement is going to witness platforms achieving a fixed set of principles by making sure the right selection, integrity, and protective measures are in place while also giving rise to better transparency over several regions around the globe. We’re now hearing about how both Microsoft and Google are also signing up to achieve the same target.

The IFES added how such regulations will give rise to a new framework for better engagement as well as support between election officials and tech firms working during that time. They would help in making the most clear policies and processes to have data shared around election time and also help in making sure voters in new democracies get the best quality of data out there today.

This will similarly enhance the communications amongst various channels as well as the election authorities while limiting barriers to ensure there’s enough support for the most authentic election period, even if the company is less known and small in scale than the rest.

Every region will feature a new set of rules surrounding data sharing as well as the whole political process. It is going to be very difficult for every tech giant to come forward with universal parameters that ensure the right protection is in place across various nations.

This latest agreement is designed to better the entire shared initiative so voters are guarded at all times while also giving rise to more knowledge shares and cooperation on such fronts between the leading providers today.

As mentioned by Meta on this front, they feel the guidelines are built on the efforts made to ensure shared practices are used that not only benefit the company but other arch rivals every year when elections are arising all over the globe.

The latest set of rules was designed at this year’s Summit for Democracy which is taking place in the Korean city of Seoul now. The latest use of tech has been a key focus for that conference and this is where global leaders will share approaches to various elements affecting the bigger and broader ecosystem for data.

This certainly seems like a bigger step in the right direction to making sure this year’s election results are as authentic as can be with Meta preparing for what will come next in terms of a mighty US election race.

Image: DIW-Aigen

Read next: Meta Rolls Out New Offer To Reduce Facebook and Instagram Subscription Fee By Half
by Dr. Hura Anwar via Digital Information World

Meta Rolls Out New Offer To Reduce Facebook and Instagram Subscription Fee By Half

Meta Platforms is giving users a chance to breathe a sigh of relief when subscribing to its Instagram and Facebook apps.

The company announced today that the monthly subscription fee was nearly half, going down from 9.9 euros to just 5.9. This was confirmed through a blog post generated by one leading Meta executive who says the decision was taken due to both privacy and antitrust concerns.

The reduction in pricing arrives at a time when the tech giant is already facing a huge amount of criticism about its rollout of subscriptions for Meta’s apps without ads. Critics say the only way to achieve this is by making users roll out fees to have privacy in check and they don’t think it’s fair to charge such a huge amount.

The rule was rolled out by Meta last year in November in regards to the Digital Markets Act that is designed to prevent personalized ads targeting users daily while they use the platforms. And that too, without any kind of consent provided. And it’s impacting their revenue source big time.

The firm mentioned how the fee model is designed to attain a balance to the growing demands of the privacy laws across the EU and the Digital Markets Act. Facebook’s parent firm also mentioned how they wished to really speed up the process for a while now because they wished to get stability.

Now, they’re dropping prices for both usual and additional accounts.

Meta’s lawyer has also chimed in to add how the company has been wanting to make the change for a while now and the fact that it’s already happening is a miracle in itself, he explained during a recent hearing for the company.

This has to be the lowest price that a Meta subscriber should pay, Meta reiterated, hoping the decision is welcomed by the masses with open arms. It makes sense as the uncertainty regarding the regulatory ordeal is present today and must settle down fast, it hopes.

However, privacy activists feel the issue here is not linked to costs. It’s more related to giving people their rights for free and not being forced to pay money to opt out of ad-targeting practices.

As per the GDPR, this is a consent that every individual is entitled to and it’s boldly documented in the privacy law of the EU.

Others feel even the fee is not the issue but the concerns arising here have to do with the approach of pay or okay. The whole initiative critics feel is to get people to press on okay, even if it’s not the free and genuine choice of them all. They don’t feel that a mere change in pricing would help in normalizing the act.

The hearing has given users of Meta and those belonging to third parties the chance to get more clarity on the subject and how it would comply with the respective DMA law.

We’ve also heard how Meta rolled out the offer of subscription fee reduction to regulators during this year’s start and is now talking about it with data protectors too, including the mighty watchdog in Ireland.

Those giving out consent for tracking will get free services funded with the help of ad revenue. But we must also remember how firms breaching the Digital Markets Act now would be liable to pay as much as 10% of the yearly global turnover.

Image: DIW-AIgen

Read next: Unedited AI Text is Appearing in Scientific Journals
by Dr. Hura Anwar via Digital Information World

Unedited AI Text is Appearing in Scientific Journals

The prevalence of AI generated text online has caused concern because of the fact that this is the sort of thing that could potentially end up lowering the bar for written content. It’s not easier than ever to churn out subpar content full of misinformation, but one might assume that this type of writing would never make its way to a scientific journal. In spite of the fact that this is the case, it turns out that that’s exactly what ended up happening with all things having been considered and taken into account.

Google Scholar, the research tool that allows you to pull up research papers, was recently found to contain phrases such as “As of my last knowledge update”. It is important to note that this phrase is famously used by ChatGPT to provide context about any and all answers it provides.

It gets worse. Apparently if you search "as of my last knowledge update" or "i don't have access to real-time data" on Google Scholar, tons of AI generated papers pop up. This is truly the worst timeline. pic.twitter.com/YXZziarUSm

— Life After My Ph.D. (@LifeAfterMyPhD) March 18, 2024

A paper called Quantum Entanglement: Examining its Nature and Implications that was published in the “Journal of Material Sciences & Manfacturing Research” contains this phrase. It bears mentioning that the previous typo is intentional, since the title of this supposedly peer reviewed journal contains it. A recent tweet also went viral after it showed the phrase popping up in scientific journals.

The combo of predatory journals, the pressure to publish, and ChatGpt is going to accelerate the destruction of the trustworthiness scientific articles.

These authors didn't even edit the AI output and the journal published it. 😭 pic.twitter.com/n7XbG5lbxh

— Koltai.bsky.social (@KolinaKoltai) March 14, 2024

However, these journals are not exactly the most renowned in the world. Rather, they were small scale journals that didn’t really have much of a reputation. This just goes to show how important it is to make sure that you are only paying attention to information provided in journals that are vetted. The low level journals that AI has been found in are thought to be troublesome in the scientific community, since they publish anything for a fee and offer a certificate of publication as well.

• Also read: From Tokyo to Silicon Valley: Exploring the Powerhouses of Global Innovation

This might cause some to lose their faith in scientifically proven information that has been independently verified. With misinformation reaching such unprecedented levels, such issues may very well end up being catastrophic in the long run as well as in the short term. As a result, many are calling for more work to be done to prevent AI being misused in such a manner.

Image: DIW-Aigen

Read next: Mastering ChatGPT: The Art of Crafting Clear and Concise Instructions

And: AI Revolution: Survey Finds 67% of Businesses Leverage AI for SEO and Content
by Zia Muhammad via Digital Information World

Study Says that Using AI in Businesses Can Impact the Security of the Company as Well as Customers

Research by the University of Sunshine Coast published a paper in AI and Ethics that said the rapid use of AI by Australian companies is putting the security of many people at risk. As many companies are using AI for a lot of work, it can have serious effects on the overall environment of the workplace. These serious effects cause violation of data and loss in business because of AI manipulation. This can be intentional as well as accidental. This study also included five points that every business should check to safely use AI for their work.

A lecturer at the university, Dr. Declan Humphreys said that there is currently a race between many companies about who will implement AI more quickly but it isn’t as easy because these companies are facing many technical issues as well as moral issues. AI models like ChatGPT, Bard, Gemini etc. use a lot of data to turn into content that seems like it is created by humans. The lecturer also added that tech firms aren’t the only ones to use AI in their works but other firms like Call Centres, Supply Chain Operators, Human Resource Management, etc also want to use AI in their work as quickly as possible.

There are a lot of talks about AI taking the jobs of people but companies aren’t ready to talk about how huge cyber security issues can be if they work with AI. If they will over use AI in their companies, it means that they are over trusting and completely relying on AI and that is not right approach. Many companies are also making their own AI models or using third-party systems without considering that their data can be hacked. AI can also make bad business decisions that can impact the overall business.

The first point in the five point checklist that the study provided if any business wants to work with AI is that the AI model you are working with should be secure and ethical. Other points include trusted and fair data collection with AI, secure data storage, maintenance of AI using ethical methods and upskilling, training and managing staff properly. Dr. Humphreys said that if any business is using AI, its top priority should be privacy and security of the customers.

Image: DIW-AI-gen

Read next: Rival IQ Report: Low Brand Engagement Across Social Media Platforms
by Arooj Ahmed via Digital Information World

Rival IQ Report: Low Brand Engagement Across Social Media Platforms

Social media has been sending fewer visitors to websites across many industries lately. This drop makes it hard for brands to know if their content is doing well or if the social platforms themselves are making it harder to reach and engage people.

A new report from Rival IQ offers insights into this issue. The report looks at how brands are doing on major social media apps. It uses data from over 5 million posts and more than 10 billion interactions, like likes and comments, on Facebook, Instagram, TikTok, and Twitter. The report gives a general view as well as details for specific industries.

For Facebook, the report shows that the average engagement rate for brands is 0.063%. This is a tiny increase from last year's 0.060%. Sports, influencers, higher education, and alcohol brands do better than others on Facebook. However, even the top performers only see a small part of their audience interacting with their posts. This suggests that looking at how far posts reach might be a better way to measure success, as most people on social media watch but don't interact.

The report questions if social media is as effective for marketing as people think. It suggests that brands might need to focus on other goals, like sales or getting people to sign up for emails, to really see how they are doing.

Instagram's average engagement rate has gone down to 0.43% from last year, continuing a downward trend. The competition is increasing, and changes in how users consume content are affecting how many people see and interact with brand posts.

Twitter, now known as X, has an average engagement rate of 0.029%, a slight drop from last year. The platform's focus has shifted away from helping brands get traffic. Yet, the decline in engagement isn't as big as expected.

TikTok saw a big drop in brand engagement, with the average rate falling to 5.69% per clip. Despite this, TikTok still offers a better chance for brands to reach and engage people compared to other platforms.

Rival IQ's report shows that brand engagement is low across all major social media apps. This could mean that people are less interested in brand content or that brands are struggling to connect with their audience. The actual results for a brand will depend on how well they understand and engage their audience on each platform.

Read next: AI Revolution: Survey Finds 67% of Businesses Leverage AI for SEO and Content
by Mahrukh Shahid via Digital Information World