ByteDance, the parent company of TikTok, has dismissed an intern for what it described as “maliciously interfering” with the training of one of its AI models. The Chinese tech giant clarified that while the intern, who was part of the advertising technology team, had no experience with ByteDance’s AI Lab, some reports circulating on social media and other platforms have exaggerated the incident’s impact.
ByteDance stated that the interference did not disrupt its commercial operations or its large language AI models. It also denied claims that the damage exceeded $10 million or affected an AI training system powered by thousands of graphics processing units (GPUs). The company highlighted that the intern was fired in August, and it has since notified their university and relevant industry bodies.
As one of the leading tech firms in AI development, ByteDance operates popular platforms like TikTok and Douyin. The company continues to invest heavily in AI, with applications including its Doubao chatbot and a text-to-video tool named Jimeng.
A London-based company, Synthesia, known for its lifelike AI video technology, is under scrutiny after its avatars were used in deepfake videos promoting authoritarian regimes. These AI-generated videos, featuring people like Mark Torres and Connor Yeates, falsely showed their likenesses endorsing the military leader of Burkina Faso, causing distress to the models involved. Despite the company’s claims of strengthened content moderation, many affected models were unaware of their image’s misuse until journalists informed them.
In 2022, actors like Torres and Yeates were hired to participate in Synthesia’s AI model shoots for corporate projects. They later discovered their avatars had been used in political propaganda, which they had not consented to. This caused emotional distress, as they feared personal and professional damage from the fake videos. Despite Synthesia’s efforts to ban accounts using its technology for such purposes, the harmful content spread online, including on platforms like Facebook.
UK-based Synthesia has expressed regret, stating it will continue to improve its processes. However, the long-term impact on the actors remains, with some questioning the lack of safeguards in the AI industry and warning of the dangers involved when likenesses are handed over to companies without adequate protections.
Meta’s Oversight Board has opened a public consultation on immigration-related content that may harm immigrants following two controversial cases on Facebook. The board, which operates independently but is funded by Meta, will assess whether the company’s policies sufficiently protect refugees, migrants, immigrants, and asylum seekers from severe hate speech.
The first case concerns a Facebook post made in May by a Polish far-right coalition, which used a racially offensive term. Despite the post accumulating over 150,000 views, 400 shares, and receiving 15 hate speech reports from users, Meta chose to keep it up following a human review. The second case involves a June post from a German Facebook page that included an image expressing hostility toward immigrants. Meta also upheld its decision to leave this post online after review.
Following the Oversight Board’s intervention, Meta’s experts reviewed both cases again but upheld the initial decisions. Helle Thorning-Schmidt, co-chair of the board, stated that these cases are critical in determining if Meta’s policies are effective and sufficient in addressing harmful content on its platform.
Sam Altman’s cryptocurrency project, Worldcoin, has rebranded as World Network and is intensifying its efforts to scan irises worldwide using its “orb” devices. The project’s core feature, World ID, acts as a digital passport to verify individuals as real humans, helping to distinguish them from AI bots online. At an event in San Francisco, World Network revealed an updated version of its orb device, boasting 5G connectivity and enhanced privacy features, alongside new initiatives to improve access to the technology.
Despite signing up over 6.9 million people since its launch in July 2023, the project has faced criticism from privacy advocates regarding the collection and storage of personal data. Several countries, including Spain and Portugal, have temporarily banned the use of the orb devices, while Argentina and Britain are currently reviewing the project.
These artificial avatars would operate on social media and online platforms, featuring realistic expressions and high-quality images akin to government IDs. JSOC also seeks technologies to produce convincing facial and background videos, including ‘selfie videos’, to avoid detection by social media algorithms.
US state agencies have previously announced frameworks to combat foreign information manipulation, citing national security threats from these technologies. Despite recognising the global dangers posed by deepfakes, SOCOM’s initiative underscores a willingness to engage with the technology for potential military advantage.
Experts expressed concern over the ethical implications and potential for increased misinformation, warning of the entirely deceptive nature of deepfakes, with no legitimate applications beyond deceit, possibly encouraging further global misuse. Furthermore, such practices pose the risk of diminished public trust in government communications, exacerbated by perceived hypocrisy in deploying such technology.
Why does it matter?
This plan reflects an ongoing interest in leveraging digital manipulation for military purposes, despite previous incidents where platforms like Meta dismantled similar US-linked networks. It further shows a contradiction in the US’s stance on deepfake use, as it simultaneously condemns similar actions by countries like Russia and China.
OpenAI has released the ChatGPT app for Windows, which is now available via the Microsoft Store. Like the Mac version launched earlier this year, it offers quick access to the AI-powered chatbot, allowing users to integrate AI into their daily activities.
The app is in an early release stage and limited to paid users, including Plus, Team, Enterprise, and Edu subscribers. A broader rollout to free-tier users is expected within the next few weeks or months. Some Mac and web version features are not yet included but are planned for future updates.
Key features missing from the Windows version include advanced voice modes, integration with Google Drive and Microsoft OneDrive, and external authentication through GPT Builder. Users, however, can still upload files and photos using the newly introduced o1 model for analysis.
ChatGPT offers convenience features such as the ‘Alt + Space’ shortcut, which allows users to bring it into focus when multitasking. It also remembers its previous position on the screen, resetting to the centre upon reopening.
X (formerly Twitter), has updated its terms of service, requiring users to file any lawsuits against the company in Texas’ Northern District in the US, a court known for conservative rulings. This change, effective November 15, appears to align with Musk’s increasing support for conservative causes, including backing Donald Trump’s 2024 presidential campaign. Critics argue the move is an attempt to ‘judge-shop,’ as the Northern District has become a popular destination for right-leaning litigants seeking to block parts of President Biden’s agenda.
X’s headquarters are in Bastrop, Texas, located in the Western District, but the company has chosen the Northern District for legal disputes. This district already hosts two lawsuits filed by X, including one against Media Matters after the watchdog group published a report linking ads on the platform to posts promoting Nazism. The move to steer legal cases to this specific court highlights the company’s efforts to benefit from a legal environment more favorable to conservative causes.
The investigation particularly focuses on two contentious posts. The first is a meme from a page linked to Poland’s far-right Confederation party, featuring former prime minister Donald Tusk in a racially charged image that alludes to the EU’s immigration pact. The image utilises language perceived as a racial slur in Poland, raising ethical concerns about its impact. The second case involves an AI-generated image posted on a German Facebook page opposing leftist and green parties. It portrays a woman with Aryan features in a stop gesture with accompanying text condemning immigrants as ‘gang-rape specialists,’ a narrative linked to perceived outcomes of the Green Party’s immigration policies. This portrayal not only uses inflammatory rhetoric but also touches on deeply sensitive cultural issues within Germany.
Thorning-Schmidt highlighted the importance of examining Meta’s current approach to managing ‘coded speech’—subtle language or imagery that carries derogatory implications while avoiding direct violations of community standards.
The board’s investigation will assess whether Meta’s policies on hate speech are robust enough to protect individuals and communities at risk of discrimination, while still allowing for critical discourse on immigration matters. Meta’s policy is designed to protect refugees, migrants, immigrants, and asylum seekers from severe attacks while allowing critique of immigration laws.
Why does it matter?
The outcome of this investigation could prompt significant changes in how Meta moderates content on sensitive topics like immigration, striking a balance between curbing hate speech and preserving freedom of expression. Moreover, Meta’s oversight board tackling politically sensitive posts shows the broader challenges social media platforms face in moderating content that balances the fine line between free expression and inciting division. It highlights the ongoing debate on the role of these platforms in managing nuanced or politically sensitive content, potentially setting a precedent.
The Cybersecurity Association of China (CSAC) has urged a security review of Intel’s products in China, alleging that the US chipmaker poses a national security risk. Although CSAC is an industry group, it has strong connections to the Chinese government, and its claims may prompt action from the Cyberspace Administration of China (CAC).
CSAC’s post on WeChat accuses Intel’s chips, including its Xeon processors used for AI, of containing vulnerabilities and backdoors allegedly tied to the US NSA. The group warns that using Intel products threatens China’s national security and critical infrastructure.
This recommendation comes amid growing US-China tensions over technology and trade. Last year, the CAC banned Chinese infrastructure operators from using products from Micron Technology after a security review, raising concerns that Intel could face a similar outcome.
Intel’s China unit responded, emphasising its commitment to product safety and quality. The company stated on its WeChat account that it will cooperate with authorities to clarify concerns. If the CAC carries out a security review, it could impact Intel’s sales in its significant Chinese market. Intel’s shares recently dropped 2.7% in US premarket trading.
The European Commission has determined that X, Elon Musk’s social media platform, does not qualify as a ‘gatekeeper’ under the Digital Markets Act (DMA), exempting it from additional compliance obligations. The Commission’s decision follows a May investigation initiated after X asserted it was not a key intermediary between businesses and consumers. While X meets user thresholds and turnover criteria, the Commission clarified that it does not significantly connect business users with end consumers.
Under the DMA, which took effect in 2023, companies must have at least 45 million end users and 10,000 business users in Europe, along with an annual turnover of €7.5 billion over the last three years, to be classified as gatekeepers. Major tech firms like Google, Amazon, Apple, Meta, Microsoft, and TikTok’s parent company ByteDance have already received gatekeeper status, imposing on them strict regulations to ensure fair competition and consumer choice.
Apple has faced penalties under the DMA, with the European Commission ruling in June that its App Store practices violated the regulations. While several companies, including Apple and Meta, have appealed their gatekeeper designations, X remains unaffected by these rules for now. This decision allows X more operational flexibility compared to its competitors, although it indicates that the Commission is closely monitoring the interactions between large platforms, businesses, and consumers in the digital marketplace.
