News of the Cyber World

[kyonides]

AI Tech Not as Advanced as Promised...

Fake It Until You Make It: Companies Are Bamboozling Investors and Customers with ‘AI Washing’

As artificial intelligence continues to make headlines, some companies are being accused of “AI washing” — overstating the capabilities of their AI technology.

BBC News reports that the phenomenon of “AI washing” has been gaining attention in recent months, as more companies claim to use AI in their products and services. The term, a play on “green washing,” refers to companies making over-inflated claims about their use of AI. This can take several forms, such as claiming to use AI when using less-sophisticated computing, overstating the efficacy of AI over existing techniques, or suggesting that AI solutions are fully operational when they are not.

One high-profile example is Amazon’s “Just Walk Out” technology, which claims to use AI to enable customers at many of its Amazon Fresh and Amazon Go shops to simply pick their items and leave, with the AI using sensors to determine what they have chosen and automatically billing them. However, reports earlier this year questioned the extent of AI’s role in the system, suggesting that around 1,000 workers in India were needed to manually check almost three quarters of the transactions. Amazon denied these reports, stating that the Indian workers were simply reviewing the system.

The rise of AI washing can be attributed to several factors, including competition for funding and the desire to appear on the cutting edge. According to OpenOcean, a UK and Finland-based investment fund for new tech firms, only 10 percent of tech start-ups mentioned using AI in their pitches in 2022, but this rose to more than a quarter in 2023 and is expected to be more than a third this year. However, as Sri Ayangar, a team member at OpenOcean, notes, “a significant disparity exists between companies claiming AI capabilities, and those demonstrating tangible AI-driven results.”

The problem is compounded by the lack of a single agreed definition of AI. As Douglas Dick, UK head of emerging technology risk at KPMG, explains, “If I asked a room of people what their definition of AI is, they would all give a different answer. The term is used very broadly and loosely, without any clear point of reference. It is this ambiguity that is allowing AI washing to emerge.”

AI washing can have concerning impacts for businesses, from overpaying for technology and services to failing to meet operational objectives the AI was expected to help them achieve. For investors, it can make it harder to identify genuinely innovative companies. And for consumers, unmet expectations from products claiming to offer advanced AI-driven solutions can erode trust in start-ups doing genuinely ground-breaking work.

Spoiler Punishes Pro-Putin Artists But Not Chinese Ones

Report: Spotify Removes Russian Artists Supportive of Vladimir Putin’s War in Ukraine

The music service Spotify has deleted the pages of several Russian musicians openly supportive of President Vladimir Putin’s brutal invasion of Ukraine, according to an independent Russian media report on Thursday.

The artists removed from the world’s largest music streaming platform include Shaman, Chicherina, Polina Gagarina, Grigory Leps, Oleg Gazmanov, and a rock group called Lyube.

The European Union (EU) sanctioned Gagarina, 37, along with 68 other individuals, in June for supporting Putin’s war.

The EU cited Gagarina, a breakout star at the Eurovision Song Contest in 2015, for performing at a Moscow political rally in March to support the war and Russia’s illegal 2014 annexation of Crimea.

Canada slapped sanctions on Gagarina in 2023 for spreading Russian propaganda. Several of her concerts were canceled in the wake of her appearance at the pro-Putin political rally in March.

The EU sanctioned Shaman, a 32-year-old singer whose real name is Yaroslav Yuryevich Dronov, on Monday because he “repeatedly participated in Kremlin-organized concerts, including the Kremlin’s anniversary event for [the Ukraine] war, and given concerts in the illegally occupied regions of Ukraine.”

The EU noted that Shaman has also performed “as part of troop entertainment events for the Russian Armed Forces.”

The EU sanctioned Leps, Gazmanov, and the band Lyube in 2022 for “actions undermining or threatening the territorial integrity, sovereignty and independence of Ukraine.”

Spotify is no longer available in Russia, having suspended operations there in March 2022, shortly after Russia invaded Ukraine. Spotify also removed its content from Russian state media outlets.

The company cited Putin’s authoritarian media laws, which threaten anyone who disparages the Russian military with up to 15 years in prison, as the main reason it could no longer do business in Russia.

“Unfortunately, recently enacted legislation further restricting access to information, eliminating free expression, and criminalizing certain types of news puts the safety of Spotify’s employees and possibly even our listeners at risk,” the company said.

Spotify told the Moscow Times on Thursday that content from all of the Russian artists it removed “met the threshold” for content that “explicitly violates our content policies or local laws.”

The streaming service took note of speculation in Russian media that the pages of the artists in question were hacked and clarified that they were deliberately removed on the instructions of Spotify management.

The Moscow Times pointed out that all of the banned Putin-supporting artists can still be found, at least for the moment, on platforms such as Apple Music and YouTube.

Spotify continues to host other artists whose conduct is reprehensible or legally questionable such as Diddy, who was caught on video beating his girlfriend; R. Kelly, who was convicted of racketeering and sexually abusing his young fans; and VaVa, a popular female Chinese rap artist who openly supports the genocidal tyranny in Beijing.

AI Mimicks Al Michaels

AI Version of Al Michaels to Deliver ‘Personalized’ Recaps During Olympics

An AI version of sports commentator Al Michaels will deliver “personalized” recaps to Peacock subscribers during the 2024 Olympics.

NBC says it will use generative AI to re-create the iconic sportscaster’s voice, which will give Peacock subscribers “personalized” recaps during the 2024 Olympic games in Paris, according to a report by Hollywood Reporter.

The company reportedly announced its AI plan — which it calls “Your Daily Olympic Recap on Peacock” — at a press event at its headquarters.

Peacock streamers will have the option to submit their names and select up to three types of sports for which they want to be given highlights. After that, an AI version of Michaels’ voice will provide daily updates based on the user’s customized request.

The AI-generated voice of the sports commentator will reportedly address users by name as it provides its daily 10-minute recap, which can also include viral moments and athlete backstories, among other categories.

There are 7 million possible ways in which the playlists could be customized for the Olympics, according to an NBCUniversal executive.

Michaels, who has provided Olympics coverage on NBC since 2010, has approved the use of his voice. “When I was approached about this, I was skeptical but obviously curious,” the sportscaster said. “Then I saw a demonstration detailing what they had in mind. I said, ‘I’m in.'”

NBCUniversal’s John Jelley said a team of NBC Sports editors developed a process to “integrate, optimize and validate state-of-the-art large language model and voice synthesis technology,” according to a report by Mashable.

Michaels is not the first individual whose voice is being trained by AI, which has become an ever-increasingly powerful tool, raising eyebrows of people concerned with how far the technology will go.

As Breitbart News reported in December, the late legendary Hollywood actor Jimmy Stewart, who was known for his distinctive drawl, can now put users of the subscription-based wellness app Calm to sleep thanks to artificial intelligence.

Comrade Bing

Communist Friendly: Microsoft’s Bing Censors More Aggressively in China than Local Competitors, Study Finds

A new study by the University of Toronto’s Citizen Lab reveals that Microsoft’s Bing translation service in China employs stricter censorship measures than its Chinese counterparts, raising concerns about the tech giant’s operations in the country.

Rest of World reports that Microsoft’s Bing translation service in China has come under scrutiny following a groundbreaking study by the University of Toronto’s Citizen Lab. The research, shared exclusively with Rest of World, indicates that Bing’s censorship practices in China are more stringent than those of leading Chinese services, including Baidu Translate and Tencent Machine Translation.

The study’s findings challenge the common perception that U.S. tech companies might be more resistant to Chinese censorship demands than their local counterparts. Jeffrey Knockel, senior research associate at Citizen Lab, explained the severity of Bing’s censorship: “If you try to translate five paragraphs of text, and two sentences contain a mention of Xi, Bing’s competitors in China would delete those two sentences and translate the rest. In our testing, Bing always censors the entire output. You get a blank. It is more extreme.”

This extensive censorship extends beyond Bing’s translation service to its China-based search engine, which was found to censor more extensively than Chinese firms’ services. The implications of these practices are significant, as Knockel noted, “Microsoft’s practices harm people’s ability to communicate with an entire demographic of people.”

Microsoft has maintained a presence in China for over two decades, with its Windows operating system capturing more than 80 percent of the market share. Bing, while not the market leader, still holds a substantial 25 percent share of the search engine market in China, according to Statcounter.

The Citizen Lab study, which analyzed 10,000 unique censorship applications across five translation services, found that Bing is the only China-based translation service to consistently produce blank outputs when encountering sensitive content. In contrast, other services like Baidu, Tencent, and NetEase silently omit triggering sentences, while Alibaba displays an error message but still translates content once the user removes the sensitive text.

Benjamin Fung, a McGill University professor and expert on AI and cybersecurity technology, suggests that Microsoft’s approach may be driven by a desire to avoid Chinese government backlash. “On a technical level, censoring more is easier to achieve. They don’t have to detect exactly which phrase is sensitive. The software just has to make the binary decision: translate or not translate,” Fung explained to Rest of World.

The study also revealed that Microsoft may have recently expanded its China-based censorship. Newly discovered censored terms on Bing’s search engine include references to the September 2023 arrest of Chinese dissident Zheng Baocheng and the Bluebird Movement protests in Taiwan from the previous month.

Huawei & US Scientists

Chinese Communist-Linked Huawei Built Secret Alliance with Top U.S. Scientific Group

Bloomberg News on Tuesday published an exposé on the relationship between China’s Huawei telecom giant and Optica, a nonprofit scientific foundation based in Washington, DC.

According to documents reviewed by Bloomberg, Huawei and Optica had much deeper and stronger ties than either was willing to acknowledge in public.

Tuesday’s report was a follow-up to Bloomberg’s report in May that revealed Huawei has been “secretly funding cutting-edge research at American universities” through Optica. The Chinese Communist-linked firm turned out to be the sole funder for an annual research competition administered by Optica, a fact that was not disclosed to entrants in the competition.

Several major U.S. universities said they would not have participated in the research competition if they had known it was financed by Huawei. After the report prompted a congressional investigation, Optica returned Huawei’s funding and removed representatives of the company from the panel of judges for its research competition.

On Tuesday, Bloomberg cited “internal Optica corporate records” that said the group was far more closely linked to Huawei than previously revealed. Among other details, a whistleblower inside the company revealed that Optica CEO Elizabeth Rogan paid a visit to Huawei headquarters in April that she did not disclose, even though her trip to China that month was public knowledge and she was chronicling her journey in real-time on social media.

The same whistleblower expressed concerns that the Chinese company was using its relationship with Optica to compromise sensitive projects funded by the U.S. government, including the Defense Advanced Research Projects Agency (DARPA).

Contrary to earlier denials by Huawei and Optica, the documents viewed by Bloomberg indicated that Huawei did have a role in choosing the winners of the research competition, and it used the contest as an opportunity to develop contractual relationships with the participants. Optica representatives said Bloomberg’s analysts were misinterpreting these documents and exaggerating the role Huawei played in choosing the winners.

Bloomberg pointed to circumstantial evidence that China has used its pipeline to the American scientific community to acquire banned technology, such as Huawei cheekily rolling out “a new smartphone featuring a 7-nanometer chip whose development U.S. export controls were supposed to foil” while Commerce Secretary Gina Raimondo was visiting China last August.

Tuesday’s report uncovered another program that did not disclose its sponsorship by Huawei, the Global Environmental Measurement and Monitoring Initiative (GEMM). The co-chair of this program was a “U.S. government chemist with the National Institute of Standards and Technology.”

Optica’s alliance with Huawei appears to date back to at least 2007, five years before the House Intelligence Committee released a landmark report that identified the Chinese company as a national security threat. Some of the U.S. scientists profiled in the report also had sponsorship deals with Huawei that predated the company’s classification as a national security threat.

The general response from Huawei and Optica has been that the titanic Chinese corporation was sincerely interested in sponsoring cutting-edge research, and the American foundation took the money without realizing how alarming its connections to Huawei might appear. Optica representatives ascribed the failures of disclosure described by Bloomberg as carelessness, rather than deliberate efforts to conceal ties to a Chinese Communist entity.

Bloomberg seemed unconvinced, noting that Optica may yet be prosecuted under the False Claims Act for “causing researchers to file fraudulent information to the U.S. government regarding their sources of funds” by concealing Huawei’s involvement from them.

Arkansas Sues Chinese App

Arkansas Sues Chinese Slave-Linked Shopping App Temu: ‘Functionally Malware’

Arkansas Attorney General Tim Griffin announced on Tuesday that his office is suing the Chinese mobile phone shopping app Temu for violating fraud and personal information laws, declaring the app “functionally malware” and a threat to America.

“Temu is not an online marketplace like Amazon or Walmart. It is a data-theft business that sells goods online as a means to an end,” Griffin said in a statement, noting that Temu is “led by a cadre of former Chinese Communist Party officials, which raises significant security risks to our country and our citizens.”

The application, which has become one of the largest e-commerce companies operating in America in the span of two years, stands accused of violating the Arkansas Deceptive Trade Practices Act (ADTPA) and the Arkansas Personal Information Protection Act (PIPA).

Temu is a shopping app that offers a wide variety of extremely inexpensive products of dubious quality, shipped directly from China. It is a subsidiary of PinDuoDuo (PDD), a Chinese corporation known for its cutthroat e-commerce tactics of selling products at steep discounts to eliminate competitors, elbowing them out of the market. Like all Chinese companies, Temu is legally required to share data it gathers from customers or any other source with the Communist Party and, thus, with the People’s Liberation Army (PLA).

Temu is also implicated in China’s ongoing genocide of Uyghurs, Kazakhs, Kyrgyz people, and other ethnic Turkic peoples in occupied East Turkistan. A report by the House Select Committee on the Chinese Communist Party published in June 2023 concluded that there is an “extremely high risk” of obtaining products tainted by slave trade when shopping on Temu and fellow Chinese e-commerce app Shein.

The Arkansas lawsuit focuses on the threat of Americans’ personal data falling into the hands of the PLA through Temu, rather than the slave trade, as lawmakers have stated Temu appears to be selling slave-tainted products to Americans through a loophole in the Uyghur Forced Labor Prevention Act (UFLPA).

CyberTruck Recalled Again

Elon’s Latest EV Fail: Tesla Cybertruck Faces Two Recalls over Wiper Failure and Loose Trim

Elon Musk’s Tesla Cybertruck has encountered yet another setback as the electric vehicle manufacturer has issued two new recalls addressing windshield wiper malfunctions and potentially detaching trunk bed trim.

PCMag reports that Tesla, the electric vehicle company led by Elon Musk, is once again facing challenges with its bizarre Cybertruck. The National Highway Traffic Safety Administration (NHTSA) has reported that the vehicle’s sole front wiper may cease functioning, significantly increasing the risk of accidents due to compromised visibility in adverse weather conditions.

The wiper issue, which impacts an estimated 230 Cybertrucks out of a possible 11,600, has already been reported by several owners. One driver described the dire situation of having to “hang out the window” to see the road during rainfall when the wiper failed. The problem, attributed to an electrical issue causing “excessive electrical current” in the wiper motor, will be addressed by Tesla through free replacements. The company plans to notify affected truck owners by mid-August.
[X Post]
Concurrently, Tesla is recalling Cybertrucks manufactured between November 13, 2023, and May 26, 2024, due to a separate issue involving improperly adhered trunk bed trim. The NHTSA estimates that 113 vehicles out of a potential 11,383 may have loose trim that could detach while driving, posing a hazard to other motorists on the road. Tesla has committed to rectifying this problem at no cost to owners as well.

These latest recalls come on the heels of previous issues that have plagued the Cybertruck since its launch. Just two months ago, Tesla recalled the vehicles due to faulty accelerator pedal covers that could potentially dislodge and force the accelerator down unexpectedly. The company has also faced scrutiny for its pattern of delaying or cancelling Cybertruck deliveries shortly before recall announcements, raising questions about transparency and quality control measures.

The Cybertruck’s troubled history is further highlighted by the fact that all four of its recalls have occurred within this year alone. The first recall addressed a warning light issue that affected virtually every Tesla vehicle ever produced. Additionally, Tesla recently issued a separate recall for 125,000 vehicles due to a faulty seatbelt warning system, although this particular issue is being resolved through an over-the-air software update.

Sony, Universal, Warner Sue AI Companies

Sony, Universal, Warner Sue AI Companies, Claiming Copyright Infringement

Artificial intelligence (AI) companies Suno and Udio faced legal action Monday launched by major record labels Sony Music, Universal Music Group and Warner Records. Allegations include committing mass copyright infringement by using the labels’ recordings to train music-generating AI systems.

The New York Post reports the companies allegedly copied music without permission to teach their systems to create music that will “directly compete with, cheapen, and ultimately drown out” human artists’ work, according to federal lawsuits filed against Udio in New York and Suno in Massachusetts.

“Our technology is transformative; it is designed to generate completely new outputs, not to memorize and regurgitate pre-existing content,” Suno CEO Mikey Shulman said in a statement.

Representatives for Udio did not immediately respond to requests for comment by the outlet. The Post report further set out:

The complaints said Suno and Udio users have been able to recreate elements of songs including The Temptations’ “My Girl,” Mariah Carey’s “All I Want for Christmas Is You” and James Brown’s “I Got You (I Feel Good),” and could generate vocals that are “indistinguishable” from musicians such as Michael Jackson, Bruce Springsteen and ABBA.

The labels asked the courts to award statutory damages of up to $150,000 per song the defendants allegedly copied.

They accused Suno of copying 662 songs and Udio of copying 1,670.

Musicians around the world have described AI as a threat to creativity however Cambridge, Mass.-based Suno and New York-based Udio have raised millions in funding this year for their bespoke systems, which create music in response to user text prompts.

The labels’ complaints said the companies have been “deliberately evasive” about the material they used to train their technology, and revealing it for public scrutiny would “admit willful copyright infringement on an almost unimaginable scale.”

Meta Might Be Lying to You

Photographers: Mark Zuckerberg’s Meta’s Labels Real Pictures as AI Fakes

Meta’s recent initiative to label AI-generated images on its social media platforms has ignited a heated debate within the photography community, as many professionals claim their non-AI images are being incorrectly tagged.

TechCrunch reports that in February, Mark Zuckerberg’s Meta announced its plan to start labeling photos created with AI tools across its social networks, including Facebook, Instagram, and Threads. The company began implementing this feature in May, attaching a “Made with AI” label to certain images. However, the rollout has not been without its challenges, as numerous users and photographers have reported instances of the label being incorrectly applied to photos that were not created using AI tools.

The controversy surrounding Meta’s labeling approach has gained significant attention, with several high-profile cases coming to light. One notable example involves a photo of the Kolkata Knight Riders winning the Indian Premier League Cricket tournament, which was erroneously tagged as AI-generated. Interestingly, the label is only visible on mobile apps and not on the web version of Meta’s platforms, adding another layer of complexity to the issue.

Photographers have been particularly vocal about their concerns, arguing that simple editing techniques should not warrant the “Made with AI” label. Former White House photographer Pete Souza found himself at the center of this debate when one of his photos was tagged with the AI label. In an email to TechCrunch, Souza explained that a change in Adobe’s cropping tool, which now requires users to “flatten the image” before saving it as a JPEG, may have triggered Meta’s algorithm to attach the label.

Souza expressed his frustration, stating, “What’s annoying is that the post forced me to include the ‘Made with AI’ even though I unchecked it.” This sentiment is echoed by many other photographers who feel their work is being misrepresented by the labeling system.

Meta has been relatively tight-lipped about the specifics of their labeling process. In a February blog post, the company mentioned that it utilizes image metadata to detect and apply the label. Meta claimed to be developing “industry-leading tools that can identify invisible markers at scale,” specifically referencing the “AI generated” information in the C2PA and IPTC technical standards.

Gemini AI Offered to Teens via School Accounts

They Want Your Kids: Google Extends Access to Ultra-Woke Gemini AI to Teens via School Accounts

Google has announced the expansion of its woke Gemini AI technology to teen students using school accounts, alongside the introduction of new tools for educators. The leftist tech giant is extending its AI plans to reach teens and children whose education often includes a Google account.

TechCrunch reports that Google’s decision to bring Gemini to teen students through their school accounts marks a major step in the tech giant’s efforts to prepare the next generation for an AI-driven future. This expansion follows the earlier rollout of Gemini to teens using personal accounts, signaling Google’s attempts to make AI technology more accessible to younger users in educational settings.

The company’s rationale behind this move is clear: to equip students with the necessary skills to navigate a world where generative AI is increasingly prevalent. Google believes that exposure to Gemini will enable students to learn more confidently, benefiting from real-time feedback and enhanced learning experiences.

The company has emphasized its commitment to responsible AI implementation in educational environments. One of the key assurances provided is that data from student interactions with Gemini will not be used to train or improve Google’s AI models. This commitment addresses potential privacy concerns that often arise when introducing new technologies in educational settings.

Sure.  As if that AI monster had never created black founding fathers that never existed or even fake popes.
Now they wants to believe they won't take advantage of their future AI monopoly. Right...

Google has also implemented several safeguards to ensure the appropriate use of Gemini by teen students. These include built-in guardrails to prevent inappropriate responses, particularly those related to illegal or age-restricted content. In an effort to promote critical thinking and information literacy, Google is actively encouraging teens to utilize the double-check feature within Gemini.

The rollout of Gemini to school accounts will be available in English across more than 100 countries. It’s important to note that the feature will be turned off by default for teen users, giving educational administrators the control to activate it as they see fit. This approach allows for a measured and controlled introduction of the technology in school environments.

Google is also enhancing its suite of tools for educators. New features include improved capabilities for creating, managing, and sharing interactive lessons. Additionally, teachers will now have the ability to manually mark assignments as missing or complete and perform bulk scoring actions, streamlining administrative tasks and allowing more time for actual teaching.

Google’s Gemini AI made big news earlier this year when the ultra-woke artificial intelligence ran amok rewriting history. Illustrating AI’s threat to students, Google’s system generated everything from black founding fathers to female indian popes.