YouTube and Those Kiddos. Greed or Weird Fascination?
September 26, 2023
![Vea4_thumb_thumb_thumb_thumb_thumb_t[1]](http://arnoldit.com/wordpress/wp-content/uploads/2023/09/Vea4_thumb_thumb_thumb_thumb_thumb_t1.gif "Vea4_thumb_thumb_thumb_thumb_thumb_t[1]")
Note: This essay is the work of a real and still-alive dinobaby. No smart software involved, just a dumb humanoid.
Google and its YouTube subsidiary are in probably in trouble again because they are spying on children. Vox explores if “Is YouTube Tracking Your Kids Again?” and sending them targeted ads. Two reports find that YouTube continues to collect data on kids despite promises not to do so. If YouTube is collecting data and sending targeted ads to young viewers it would violate the Children’s Online Privacy and Protection act (COPPA) and Google’s consent decree with the FTC.
Google agreed to the consent decree with the FTC to stop collecting kids’ online activity and selling it to advertisers. In order to regulate and comply with the decree and COPPA, YouTube creators must say if their channels or individual videos are kid friendly. If they are designated kid friendly then Google doesn’t collect data on the viewers. This only occurs on regular YouTube and not YouTube Kids.
Fairplay and Analytics researched YouTube data collection and released compromising reports. Fairplay, a children’s online safety group, had an ad campaign on YouTube and asked for it to target made for kids videos. The group discovered their ads played on videos that were kids only, basically confirming that targeted ads are still being shown to kids. Analytics found evidence that supports kid data collection too:
“The firm found trackers that Google uses specifically for advertising purposes and what appear to be targeted ads on “made for kids” videos. Clicking on those ads often took viewers to outside websites that definitely did collect data on them, even if Google didn’t. The report is careful to say that the advertising cookies might not be used for personalized advertising — only Google knows that — and so may still be compliant with the law. And Adalytics says the report is not definitively saying that Google violated COPPA: ‘The study is meant to be viewed as a highly preliminary observational analysis of publicly available information and empirical data.’”
Google denies the allegations and claims the information in the reports are skewed. YouTube states that ads on made for kids videos are contextual rather than targeted, implying they are shown to all kids instead of individualizing content. If Google and YouTube are to be in violation of the FTC decree and COPPA, Alphabet Inc would pay a very expensive fine.
It is hard to define what services and products that Google can appropriately offer kids. Google has a huge education initiative with everything from laptops to email services. Republicans and Democrats agree that it is important to protect kids online and hold Google and other companies liable. Will Google pay fines and not worry about the consequences? I have an idea. Let’s ask Meta’s new kid-oriented AI initiative. That sounds like a fine idea.
Whitney Grace, September 26, 2023
Amazon Switches To AI Review Summaries
September 22, 2023
The online yard sale eBay offers an AI-generated description feature for sellers. Following in the same vein, Engadget reports that, “Amazon Begins Rolling Out AI-Generated Review Summaries” for products with clickable keywords. Amazon announced in June 2023 that it was testing an AI summary tool across a a range of products. The company officially launched the tool in August declaring that AI is at the heart of Amazon.
Amazon developed the AI summary tool so consumers can read buyers’ opinions without scrolling through pages of information. The summaries are described as a wrap-up of customer consensus akin to film blurbs on Rotten Tomatoes. The AI summaries contain clickable tags that showcase common words and consistent themes from reviews. Clicking on the tags will take consumers to the full review with the information.
AI-generated review summaries bring up another controversial topic: Amazon and fake reviews. Fake reviews litter the selling platform like a slew of counterfeit products Amazon, eBay, and other online selling platforms battle. While Amazon claims it takes a proactive stance to detect and delete the reviews, it does not catch all the fakes. It is speculated that AI-generated reviews from ChatGPT or other chatbots are harder for Amazon to catch.
In regards to using its own AI summary tool, Amazon plans to only use it on verified purchases and using more AI models to detect fake reviews. Humans will be used for clarification with their more discerning organic brains. Amazon said about its news tool:
“‘We continue to invest significant resources to proactively stop fake reviews,’ Amazon Community Shopping Director Vaughn Schermerhorn said. ‘This includes machine learning models that analyze thousands of data points to detect risk, including relations to other accounts, sign-in activity, review history, and other indications of unusual behavior, as well as expert investigators that use sophisticated fraud-detection tools to analyze and prevent fake reviews from ever appearing in our store. The new AI-generated review highlights use only our trusted review corpus from verified purchases, ensuring that customers can easily understand the community’s opinions at a glance.’”
AI tools are trained using language models that contain known qualitative errors. The same AI tools are used to teach more AI and so on. While we do not know what Amazon is using to train its AI summary tool, we would not be surprised if the fake reviews are using similar training models to Amazon’s. It will come down to Amazon AI vs. counterfeit AI. Who will win?
Whitney Grace, September 22, 2023
Kill Off the Dinobabies and Get Younger, Bean Counter-Pleasing Workers. Sound Familiar?
September 21, 2023
Note: This essay is the work of a real and still-alive dinobaby. No smart software involved, just a dumb humanoid.
I read “Google, Meta, Amazon Hiring low-Paid H1B Workers after US Layoffs: Report.” Is it accurate? Who knows? In the midst of a writers’ strike in Hollywood, I thought immediately about endless sequels to films like “Batman 3: Deleting Robin” and Halloween 8: The Night of the Dinobaby Purge.”
The write up reports a management method similar to those implemented when the high school science club was told that a school field trip to the morgue was turned down. The school’s boiler suffered a mysterious malfunction and school was dismissed for a day. Heh heh heh.
I noted this passage:
Even as global tech giants are carrying out mass layoffs, several top Silicon Valley companies are reportedly looking to hire lower-paid tech workers from foreign countries. Google, Meta, Amazon, Microsoft, Zoom, Salesforce and Palantir have applied for thousands of H1B worker visas this year…
I heard a rumor that IBM used a similar technique. Would Big Blue replace older, highly paid employees with GenX professionals not born in the US? Of course not! The term “dinobabies” was a product of spontaneous innovation, not from a personnel professional located in a suburb of New York City. Happy bean counters indeed. Saving money with good enough work. I love the phrase “minimal viable product” for “minimally viable” work environments.
There are so many ways to allow people to find their futures elsewhere. Shelf stockers are in short supply I hear.
Stephen E Arnold, September 21, 2023
Just TikToking Along, Folks
September 21, 2023
Beleaguered in the US, its largest market, TikTok is ready to embrace new options in its Southeast Asian advance. CNBC reports, “TikTok Shop Strikes ‘Buy Now, Pay Later’ Partnership in Malaysia As Part of E-Commerce Push.” Writer Cheila Chiang reports:
“The partnership comes as TikTok looks to markets outside of the U.S. for growth. While the U.S. is the company’s largest market, TikTok faces headwinds there after Montana became the first state to ban the app. The app has also been banned in India. In recent months, TikTok Shop has been aggressively expanding into e-commerce in Southeast Asia, competing against existing players like Sea’s Shopee and Alibaba’s Lazada. TikTok’s CEO previously said the company will pour ‘billions of dollars’ into Southeast Asia over the next few years. As of April, TikTok said it has more than 325 million monthly users in Southeast Asia. In June, the company said it would invest $12.2 million to help over 120,000 small and medium-sized businesses sell online. The investment consists of cash grants, digital skills training and advertising credits for these businesses.”
What a great idea for the teenagers who are the largest cohort of TikTok users. Do they fully grasp the pay later concept and its long-term effects? Sure, no problem. Kids love to work at part time jobs, right? As long as major corporations get to expand as desired, that is apparently all that matters.
Cynthia Murrell, September 21, 2023
Big Tech: Your Money or Your Digital Life? We Are Thinking
September 20, 2023
Why is anyone surprised that big tech companies want to exploit AI for profit? Business Insider gives a quick rundown on how big tech advertised AI as beneficial research tool while now they are prioritizing it as commercial revenue tool in, “Silicon Valley Presented AI As A Noble Research Tool. Now It’s All About Cold, Hard Cash.”
Big tech companies presented AI research as a societal boon and would share the findings with everyone. The research was done without worrying about costs and it is the ideal situation or ultimate discovery. Google wrote off $1.3 million of DeepMind’s debt to demonstrate its commitment to advancing AI research.
As inflation rises, big tech companies are worried about their bottom lines. ChatGPT and similar algorithms has made significant headway in AI science, so big tech companies are eager to exploit it for money. Big tech companies are racing to commercialize chatbots by promoting the benefits with consumers. Competitors are forced to develop their own chatbots or lose business.
Meta is prioritizing AI research but ironically sacked a team researching protein folding. Meta wants to cut the fat to concentrate on profits. Unfortunately the protein folding was axed despite how understanding protein folding could help scientists understand diseases, such as Parkinson’s and Alzheimer’s.
Google is focusing on net profits too. One good example is a new DeepMind unit that shares AI research papers to improve people’s lives as well as products. Google did make the new large language model Llama 2 an open source tool for businesses with fewer than 700 million monthly active users. Google continues to output smart chatbots. Hey, students, isn’t that helpful?
It is unfortunate that humans are inherently selfish beings. If we did everything for the benefit of society it would be great, but history has shown socialism and communism does not work. There is a way to fund exploratory research without worrying about money. We just have not found it yet.
Whitney Grace, September 20, 2023
Profits Over Promises: IBM Sells Facial Recognition Tech to British Government
September 18, 2023
Just three years after it swore off any involvement in facial recognition software, IBM has made an about-face. The Verge reports, “IBM Promised to Back Off Facial Recognition—Then it Signed a $69.8 Million Contract to Provide It.” Amid the momentous Black Lives Matter protests of 2020, IBM’s Arvind Krishna wrote a letter to Congress vowing to no longer supply “general purpose” facial recognition tech. However, it appears that is exactly what the company includes within the biometrics platform it just sold to the British government. Reporter Mark Wilding writes:
“The platform will allow photos of individuals to be matched against images stored on a database — what is sometimes known as a ‘one-to-many’ matching system. In September 2020, IBM described such ‘one-to-many’ matching systems as ‘the type of facial recognition technology most likely to be used for mass surveillance, racial profiling, or other violations of human rights.'”
In the face of this lucrative contract IBM has changed its tune. It now insists one-to-many matching tech does not count as “general purpose” since the intention here is to use it within a narrow scope. But scopes have a nasty habit of widening to fit the available tech. The write-up continues:
“Matt Mahmoudi, PhD, tech researcher at Amnesty International, said: ‘The research across the globe is clear; there is no application of one-to-many facial recognition that is compatible with human rights law, and companies — including IBM — must therefore cease its sale, and honor their earlier statements to sunset these tools, even and especially in the context of law and immigration enforcement where the rights implications are compounding.’ Police use of facial recognition has been linked to wrongful arrests in the US and has been challenged in the UK courts. In 2019, an independent report on the London Metropolitan Police Service’s use of live facial recognition found there was no ‘explicit legal basis’ for the force’s use of the technology and raised concerns that it may have breached human rights law. In August of the following year, the UK’s Court of Appeal ruled that South Wales Police’s use of facial recognition technology breached privacy rights and broke equality laws.”
Wilding notes other companies similarly promised to renounce facial recognition technology in 2020, including Amazon and Microsoft. Will governments also be able to entice them into breaking their vows with tantalizing offers?
Cynthia Murrell, September 18, 2023
Microsoft: Good Enough Just Is Not
September 18, 2023
Was it the Russian hackers? What about the special Chinese department of bad actors? Was it independent criminals eager to impose ransomware on hapless business customers?
No. No. And no.
The manager points his finger at the intern working the graveyard shift and says, “You did this. You are probably worse than those 1,000 Russian hackers orchestrated by the FSB to attack our beloved software. You are a loser.” The intern is embarrassed. Thanks, Mom MJ. You have the hands almost correct… after nine months or so. Gradient descent is your middle name.
“Microsoft Admits Slim Staff and Broken Automation Contributed to Azure Outage” presents an interesting interpretation of another Azure misstep. The report asserts:
Microsoft’s preliminary analysis of an incident that took out its Australia East cloud region last week – and which appears also to have caused trouble for Oracle – attributes the incident in part to insufficient staff numbers on site, slowing recovery efforts.
But not really. The report adds:
The software colossus has blamed the incident on “a utility power sag [that] tripped a subset of the cooling units offline in one datacenter, within one of the Availability Zones.”
Ah, ha. Is the finger of blame like a heat seeking missile. By golly, it will find something like a hair dryer, fireworks at a wedding where such events are customary, or a passenger aircraft. A great high-tech manager will say, “Oops. Not our fault.”
The Register’s write up points out:
But the document [an official explanation of the misstep] also notes that Microsoft had just three of its own people on site on the night of the outage, and admits that was too few.
Yeah. Work from home? Vacay time? Managerial efficiency planning? Whatever.
My view of this unhappy event is:
- Poor managers making bad decisions
- A drive for efficiency instead of a drive toward excellence
- A Microsoft Bob moment.
More exciting Azure events in the future? Probably. More finger pointing? It is a management method, is it not?
Stephen E Arnold, September 18, 2023
Turn Left at Ethicsville and Go Directly to Immoraland, a New Theme Park
September 14, 2023
Note: This essay is the work of a real and still-alive dinobaby. No smart software involved, just a dumb humanoid.
Stanford University lost a true icon of scholarship. Why is this individual leaving the august institution, a hot spot of modern ethical and moral discourse. Yeah, the leader apparently confused real and verifiable data with less real and tough-to-verify data. Across the country, an ethics professor no less is on leave or parked in an academic rest area over a similar allegation. I will not dwell on the outstanding concept of just using synthetic data to inform decision models, a practice once held in esteem at the Stanford Artificial Intelligence Lab.
“Gasp,” one PhD utters. An audience of scholars reveals shock and maybe horror when a colleague explains that making up, recycling, or discarding data at odds with the “real” data is perfectly reasonable. The brass ring of tenure and maybe a prestigious award for research justify a more hippy dippy approach to accuracy. And what about grants? Absolutely. Money allows top-quality research to be done by graduate assistants. Everyone needs someone to blame. MidJourney, keep on slidin’ down that gradient descent, please.
“Scientist Shocks Peers by Tailoring Climate Study” provides more color for these no-ethics actions by leaders of impressionable youth. I noted this passage:
While supporters applauded Patrick T. Brown for flagging what he called a one-sided climate “narrative” in academic publishing, his move surprised at least one of his co-authors—and angered the editors of leading journal Nature. “I left out the full truth to get my climate change paper published,” read the headline to an article signed by Brown…
Ah, the greater good logic.
The write up continued:
A number of tweets applauded Brown for his “bravery”, “openness” and “transparency”. Others said his move raised ethical questions.
The write up raised just one question I would like answered: “Where has education gone?” Answer: Immoraland, a theme park with installations at Stanford and Harvard with more planned.
Stephen E Arnold, September 14, 2023
What Is More Important? Access to Information or Money
September 14, 2023
Laws that regulate technology can be outdated because they were written before the technology was invented. While that is true, politicians have updated laws to address situations that arise from advancing technology. Artificial intelligence is causing a flurry of new legislative concerns. The Conversation explains that there are already laws regulating AI on the books but they are not being follow: “Do We Need A New Law For AI? Sure-But First We Could Try Enforcing The Laws We Already Have.”
In the early days of the Internet and mass implementation of computers, regulation was a bad work akin to censoring freedom of speech and would also impede technology progress. AI technology is changing that idea. Australian Minister for Industry and Science Ed Husic is leading the charge for an end to technology self-regulation that could inspire lawmakers in other countries.
Husic wants his policies to focus on high risk issues related to AI and balancing the relationship between humans and machines. He no longer wants the Internet and technology to be a lawless wild west. Big tech leaders such as OpenAI Chief Executive Sam Altman said regulating AI was essential. OpenAI developed the ChatGPT chatbot/AI assistant. Altman’s statement comes ten years after Facebook founder Mark Zuckerberg advised people in the tech industry to move fast and break things. Why are tech giants suddenly changing their tune?
One idea is that tech giants understand the dangers associated with unbridled AI. They realize without proper regulation, AI’s negative consequences could outweigh the positives.
There are already AI regulating laws in most countries but it refers to technology in general:
“Our current laws make clear that no matter what form of technology is used, you cannot engage in deceptive or negligent behavior.
Say you advise people on choosing the best health insurance policy, for example. It doesn’t matter whether you base your advice on an abacus or the most sophisticated form of AI, it’s equally unlawful to take secret commissions or provide negligent advice.”
The article was written by tech leaders at the Human Technology Institute located at the University of Technology Sydney, who are calling for Australia to create a new government role, the AI Commissioner. This new role would be an independent expert advisor to the private and government sector to advise businesses and lawmakers on how to use and enforce AI within Australia’s laws. Compared to North America, the European Union, and many Asian countries, Australia has dragged its heels developing AI laws.
The authors stress that personal privacy must be protected like the laws that already exist in Europe. Also they cite examples of how mass-automation of tasks led to discrimination and bureaucratic nightmares.
An AI Commissioner is a brilliant idea but it places the responsibility on one person. A small, regulating board monitored like other government bodies would be a better idea. Since the idea is logical the Australian government will fail to implement it. That is not a dig on Australia. Any and all governments fail at implementing logical plans.
Whitney Grace, September 14, 2023
YouTube and Click Fraud: A Warning Light Flashing?
September 13, 2023
Note: This essay is the work of a real and still-alive dinobaby. No smart software involved, just a dumb humanoid.
I spotted a link to a 16 minute YouTube long form, old-fashioned video from Lon.TV titled YouTube Invalid Traffic. The person who does Lon.TV usually reviews gadgets, but this video identifies a demonetization procedure apparently used by the non-monopoly Google. (Of course, I believe Google’s assertion that almost everyone uses Google because it is just better.)
The creator reads an explanation of an administrative action and says, “What does this mean?” Would a non-monopoly provide a non explanation? Probably a non not. Thanks, MidJourney, the quality continues to slip. Great work.
Lon.TV explains that the channel received a notice of fraudulent clicks. The “fix”, which YouTube seems to implement unilaterally and without warning, decreases a YouTuber’s income. The normal Google “help” process results in words which do not explain the details of the Google-identified problem.
Click fraud has been a tricky issue for ad-supported Google for many years. About a decade ago, a conference organizer wanted me to do a talk about click fraud, a topic I did not address in my three Google monographs. The reports for a commercial company footing the bill for my research did get information about click fraud. My attorney at the time (may he rest in peace) advised me to omit that information from the monographs published by a now defunct publisher in the UK. I am no legal eagle, but I do listen to them, particularly when it costs me several hundred dollars an hour.
Click fraud is pretty simple. One can have a human click on a link.l If one is serious, one can enlist a bunch of humans using an ad in Craigslist.com. A more enterprising click fraud player would write a script and blast through a target’s ad budget, rack up lots of popularity points, or make a so-so video into the hottest sauce pan on the camp fire.
Lon.TV’s point is that most of his site’s traffic originates from Google searches. A person looking for a camera review runs a query on Google. The Google results point to a Lon.TV video. The person clicks on the Google generated link, and the video plays. The non-monopoly explains, as I understand it, that the fraudulent clicks are the fault of the YouTuber. So, the bad actor is the gadget guy at Lon.TV.
I think there is some useful information or signals in this video. I shall share my observations:
- Click fraud, based on my research a decade ago, was indeed a problem for the non-monopoly. In fact, the estimable company was trying to figure out how to identify fraudulent clicks and block them. The work was not a path to glory, so turnover often plagued those charged with stamping out click fraud. Plus, the problem was “hard.” Simple fixes like identifying lots of clicks in a short time were easily circumvented. More sophisticated ones like figuring out blocks of IP addresses responsible for lots of time spaced clicks were okay until the fraudsters figured out another approach. Thus, cat-and-mouse games began.
- The entire point of YouTube.com is to attract traffic. Therefore, it is important to recognize what is a valid new trend like videos of females wearing transparent clothing is recognized and clicks on dull stuff like streaming videos of a view of an erupting volcano are less magnetic. With more clicks, many algorithmic beavers jump in the river. More clicks means more ads pushed. The more ads pushed means more clicks on those ads and, hence, more money. It does not take much thought to figure out that a tension exists between lots of clicks Googlers and block those clicks Googlers. In short, progress is slow and money generation wins.
- TikTok has caused Google to undermine its long form videos to deal with the threat of the China-linked competitor. The result has been an erosion of clicks because one cannot inject as many ads into short videos as big boy videos. Oh, oh. Revenue gradient decline. Bad. Quick fix. Legitimize keeping more ad revenue? Would a non monopoly do that?
- The signals emitted by Lon.TV indicate that Google’s policy identified by the gadget guy is to blame the creator. Like many of Google’s psycho-cognitive methods used to shift blame, the hapless creator is punished for the alleged false clicks. The tactic works well because what’s the creator supposed to do? Explain the problem in a video which is not pushed?
Net net: Click fraud is a perfect cover to demonetized certain videos. What happens to the ad money? Does Google return it to the advertiser? Does Google keep it? Does Google credit the money back to the advertiser’s account and add a modest “handling fee”? I don’t know, and I am pretty sure the Lon.TV fellow does not either. Furthermore, I am not sure Google “knows” what its different units are doing about click fraud. What’s a non-monopoly supposed to do? I think the answer is, “Make money.” More of these methods are likely to surface in the future.
Stephen E Arnold, September 13, 2023
-
- Subscribe to Beyond Search
Feature archive
News archive
-
