New research by The Guardian reveals that ChatGPT Search, OpenAI’s recently launched AI-powered search tool, can be misled into generating false or overly positive summaries. By embedding hidden text in web pages, researchers demonstrated that the AI could ignore negative reviews or even produce malicious code.
The feature, designed to streamline browsing by summarising content such as product reviews, is susceptible to hidden text attacks—a well-known vulnerability in large language models. While this issue has been studied before, this marks the first time such manipulation has been proven on a live AI search tool.
OpenAI did not comment on this specific case but stated it employs measures to block malicious websites and is working to improve its defences. Experts note that competitors like Google, with more experience in search technology, have developed stronger safeguards against similar threats.
Google is introducing major updates to its Maps Timeline feature, shifting towards enhanced privacy by storing location history locally on users’ devices. This change means that any unbacked location history data will be automatically deleted after three months unless users take action to save it. Notifications regarding the update will provide a six-month window to back up data, urging users to preserve personal memories or valuable travel records before permanent deletion.
To save Timeline data, users can utilise Google’s export tools on their devices or through the web. This process allows users to maintain access to their past location history while ensuring it remains private and under their control. However, some users have reported unexpected deletions, emphasising the importance of regular backups to avoid data loss.
These updates are part of the US-based tech giant’s broader efforts to address privacy concerns by limiting the storage of sensitive data on its servers. While this move enhances data security, users are encouraged to take proactive steps to back up important information, explore privacy features like incognito mode, and consider additional tools such as VPNs to protect their online activity.
Vietnam’s new internet law, known as ‘Decree 147,’ came into effect Wednesday, requiring platforms like Facebook and TikTok to verify user identities and share data with authorities upon request. Critics view the move as a crackdown on freedom of expression, with activists warning it will stifle dissent and blur the lines between legal and illegal online activity. Under the rules, tech companies must store verified information alongside users’ names and dates of birth and remove government-designated “illegal” content within 24 hours.
The decree also impacts the booming social commerce sector by allowing only verified accounts to livestream. Additionally, it imposes restrictions on gaming for minors, limiting sessions to one hour and a maximum of 180 minutes daily. Vietnam, with over 65 million Facebook users and a growing gaming population, may see significant disruptions in online behaviour and businesses.
Critics liken the law to China’s tight internet controls. Activists and content creators have expressed fear of persecution, citing recent examples like the 12-year prison sentence for a YouTuber critical of the government. Despite the sweeping measures, some local businesses and gamers remain sceptical about enforcement, suggesting a wait-and-see approach to the decree’s real-world impact.
Bluesky has seen explosive growth in recent months, surpassing 25 million users. The platform, which promotes decentralisation and user control, gained attention during the US elections and after X’s brief ban in Brazil. Bluesky has become an appealing alternative for those disenchanted with traditional platforms like Meta and X, offering curated features and a community-focused experience.
Rapid growth, however, has introduced significant challenges. Bots and AI-driven accounts have flooded the site, spreading misinformation and cluttering user interactions. The platform’s small team has worked swiftly to combat these issues, increasing its moderation capacity and introducing new tools to tackle impersonation and spam. Despite these efforts, the fight against AI bots and disinformation continues to grow more complex.
Bluesky’s commitment to decentralisation and user control has attracted users frustrated with larger platforms’ power dynamics. Experts caution, however, that the platform faces hurdles in maintaining its integrity while scaling its operations. Political and social fragmentation in online spaces could also limit Bluesky’s growth compared to visual-heavy platforms like TikTok and Instagram, which dominate younger audiences.
As the platform navigates its challenges, its future remains uncertain. Balancing growth, moderation, and user satisfaction will be critical to establishing Bluesky as a sustainable alternative in the competitive social media landscape.
Google contractors improving the Gemini AI model have been tasked with comparing its responses against those of Anthropic’s Claude, according to internal documents reviewed by TechCrunch. The evaluation process involves scoring responses on criteria such as truthfulness and verbosity, with contractors given up to 30 minutes per prompt to determine which model performs better. Notably, some outputs identify themselves as Claude, sparking questions about Google’s use of its competitor’s model.
Claude’s responses, known for emphasising safety, have sometimes refused to answer prompts deemed unsafe, unlike Gemini, which has faced criticism for safety violations. One such instance involved Gemini generating responses flagged for inappropriate content. Despite Google’s significant investment in Anthropic, Claude’s terms of service prohibit its use to train or build competing AI models without prior approval.
A spokesperson for Google DeepMind stated that while the company compares model outputs for evaluation purposes, it does not train Gemini using Anthropic models. Anthropic, however, declined to comment on whether Google had obtained permission to use Claude for these tests. Recent revelations also highlight contractor concerns over Gemini producing potentially inaccurate information on sensitive topics, including healthcare.
Greek authorities have made their first-ever cryptocurrency seizure, confiscating 273,000 USDT (Tether) as part of a criminal investigation. The operation, conducted in December, was carried out under the supervision of the Greek European Public Prosecutor’s Office and involved collaboration with various law enforcement departments, including the Digital Evidence Examination Department.
The seizure, which is part of the ongoing ‘Admiral’ operation, highlights the growing challenges law enforcement faces in dealing with advanced technologies like blockchain and cryptocurrencies. Cryptocurrencies, known for their anonymity and security features, are often used in criminal activities such as fraud and money laundering. Experts stress the need for precision and expertise in handling digital assets, as mistakes can lead to irreversible losses.
Crypto-related scams are becoming more common in Greece, with many victims falling prey to fraudulent schemes. As cryptocurrencies gain popularity, particularly with the rise of Bitcoin and NFTs, the lack of understanding among the public increases the risk of scams. Experts warn that technological advances in AI are making these scams harder to detect, even for experienced investors.
In addition to combating fraud, authorities are also focusing on the management of seized cryptocurrencies, with plans to convert them into funds for the state, similar to practices in other European countries.
According to state media reports, Iran has lifted its ban on Meta’s WhatsApp and Google Play, marking a tentative move toward easing internet restrictions. Known for its stringent online censorship, Iran has long restricted access to US-based platforms like Facebook, Twitter, and YouTube, though many Iranians bypass these blocks using virtual private networks.
The decision, announced after a meeting led by President Masoud Pezeshkian, reflects a ‘positive majority vote’ to restore access to some popular foreign platforms. Information and Communications Technology Minister Sattar Hashemi hailed the move as the ‘first step in removing internet limitations.’
Social media has played a significant role in Iran, particularly as a tool for organising anti-government protests. In response to such restrictions, the United States has urged Big Tech companies to support efforts to circumvent censorship in countries like Iran.
As Germany prepares for national elections on February 23, political parties are outlining their tech policy priorities, including digitalisation, AI, and platform regulation. Here’s where the leading parties stand as they finalise their programs ahead of the vote.
The centre-right CDU, currently leading in polls with 33%, proposes creating a dedicated Digital Ministry to streamline responsibilities under the Ministry of Transport. The party envisions broader use of AI and cloud technology in German industry while simplifying citizen interactions with authorities through digital accounts.
Outgoing Chancellor Olaf Scholz’s SPD, polling at 15%, focuses on reducing dependence on US and Chinese tech platforms by promoting European alternatives. The party also prioritises faster digitalisation of public administration and equitable rules for regulating AI and digital platforms, echoing EU-wide goals of tech sovereignty and security.
The Greens, with 14% support, highlight the role of AI in reducing administrative workloads amid labour shortages. They stress the need for greater interoperability across IT systems and call for an open-source strategy to modernise Germany’s digital infrastructure, warning that the country lags behind EU digitalisation targets.
The far-right AfD, projected to secure 17%, opposes EU platform regulations like the Digital Services Act and seeks to reverse Germany’s adoption of the NetzDG law. The party argues these measures infringe on free speech and calls for transparency in funding non-state actors and NGOs involved in shaping public opinion.
The parties’ contrasting visions set the stage for significant debates on the future of technology policy in Germany.
Albania has announced a one-year nationwide ban on TikTok, citing concerns about the platform’s influence on children. The decision follows the fatal stabbing of a 14-year-old boy in November, reportedly linked to social media disputes. Prime Minister Edi Rama revealed the ban as part of a broader strategy to enhance school safety after consultations with parents and teachers.
The Prime Minister has criticised TikTok and similar platforms for encouraging youth violence. Videos supporting the killing were shared online, raising alarms about the role of social media in such incidents. Rama stated that society, not children, bears responsibility for the issue, describing TikTok as a platform that holds children ‘hostage’.
TikTok responded by seeking clarity from the Albanian government, claiming no evidence linked the involved teens to the platform. A spokesperson suggested another platform might have hosted the content tied to the incident.
President-elect Donald Trump hinted at allowing TikTok to continue operating in the US, at least temporarily, citing the platform’s significant role in his presidential campaign. Speaking to conservative supporters in Phoenix, Arizona, Trump shared that his campaign content had garnered billions of views on TikTok, describing it as a “beautiful” success that made him reconsider the app’s future.
TikTok’s parent company, ByteDance, has faced pressure from US lawmakers to divest the app over national security concerns, with allegations that Chinese control of TikTok poses risks to American data. The US Supreme Court is set to decide on the matter, as ByteDance challenges a law that could force divestment. Without a favourable ruling or compliance with the law, TikTok could face a US ban by January 19, just before Trump takes office.
Trump’s openness to TikTok contrasts with bipartisan support for stricter measures against the app. While the Justice Department argues that Chinese ties to TikTok remain a security threat, TikTok counters that its user data and operations are managed within the US, with storage handled by Oracle and moderation decisions made domestically. Despite ongoing legal battles, Trump’s remarks and a recent meeting with TikTok’s CEO suggest he sees potential in maintaining the platform’s presence in the US market.
