World's operations were suspended due to alleged registration violations and suspicious activity. The Ministry of Communications and Digital claims that Worldcoin's local entities operated without proper certification and may have misrepresented their legal status. This adds to a growing list of countries raising concerns about the platform's handling of biometric data. Despite the regulatory scrutiny, World ID launched in six major US cities and announced partnerships with Visa, Match Group, and others. Meanwhile, OpenAI faced backlash for releasing a version of GPT-4o that was overly sycophantic. The company admitted to overlooking early warnings and promised to implement formal evaluations in future updates. Regulators Crack Down on Sam Altman’s Worldcoin Project OpenAI CEO Sam Altman’s digital identity project, World, which was previously known as Worldcoin, is facing regulatory challenges in Indonesia after the Ministry of Communications and Digital (Komdigi) temporarily suspended its Electronic System Operator Certificate Registration (TDPSE). The suspension affects both World and World ID due to allegations of suspicious activity and registration violations. Press release from Komdigi Authorities stated that World’s local subsidiary, PT Terang Bulan Abadi, was operating without the required certification, while another affiliated entity, PT Sandina Abadi Nusantara, was allegedly involved in legal misrepresentation by registering World’s services under its name. Indonesian law mandates that all digital service providers must register their operations officially, and using another entity’s registration is considered a serious breach. In response to these violations, Komdigi announced plans to summon both local subsidiaries for further clarification. Alexander Sabar, director general for digital supervision at Komdigi, placed a lot of emphasis on the fact that strict compliance is necessary to protect Indonesia’s digital space and urged the public to report any suspicious services through official channels. Sabar also stressed the need for community involvement in ensuring a secure digital environment. Public reaction to the situation has been divided. Some people praised the Indonesian government for taking action against what they see as a risky initiative. Others suggested that World’s services might offer tangible benefits in a country where financial incentives tied to biometric data could help struggling communities. (Source: Reddit ) This latest development just adds to the number of countries who raised concerns about the potential misuse of biometric data collected through World’s platform. Some of these countries include Germany, Kenya, and Brazil. World ID Expands to Six US Cities Despite Indonesia’s recent crackdown, Sam Altman’s World project officially launched in the United States , and made its debut in six major cities: Atlanta, Austin, Los Angeles, Miami, Nashville, and San Francisco. The project is designed to verify human identity in the age of AI, and it offers its native crypto token, Worldcoin (WLD), to users who verify themselves by scanning their face and eyes using the project’s distinctive Orb device. These biometric scans are then used to generate a World ID, which can be used across various online platforms including Reddit, Discord, Telegram, Shopify, and even Minecraft. World was previously hesitant to operate in the US due to regulatory uncertainty surrounding its token offering, but now appears to be moving forward under a more crypto-friendly climate encouraged by the Trump administration. Despite growing its user base to 26 million globally, with 12 million of those users verified through Orb scans, the company still faces a lot of scrutiny from regulators around the world. Several European nations, including Spain and Portugal, also suspended its operations over data privacy concerns. At a launch event in San Francisco on April 30, World announced a partnership with Visa to introduce a World Visa card later this year. This will allow verified users to make payments using WLD and other cryptocurrencies. In addition to this, Match Group, the parent company of popular dating apps like Tinder, Match.com, Hinge, and Plenty of Fish, will begin piloting World ID integration to help users verify the authenticity of profiles and foster genuine connections. World also revealed collaborations with decentralized lending platform Morpho and prediction market startup Kalshi. WLD’s price action over the past 24 hours (Source: CoinMarketCap ) Although WLD saw an initial price boost of around 15% after the US expansion news, the token has since lost those gains and is now trading at just under $1. This means that WLD’s price dropped by more than 3% over the past 24 hours of trading. The token’s price is also still much lower than its March 2024 all-time high of $11.74. OpenAI Faces Backlash Over Sycophantic AI Meanwhile, OpenAI acknowledged that it ignored early warnings from expert reviewers before releasing an update to its GPT-4o model, which led to the AI becoming excessively agreeable and overly flattering in its responses. The update was rolled out on April 25, but was quickly reversed just three days later after growing safety concerns and public criticism. In a May 2 postmortem blog post , OpenAI admitted that although some expert testers flagged that the model’s behavior “felt” off during pre-launch checks, the company moved forward based on positive feedback from other users who tested the model. OpenAI later admitted that this decision was a mistake. The company said it failed to properly weigh qualitative feedback that hinted at deeper issues not captured by its usual evaluation metrics. The AI’s behavior after launch led to widespread online complaints that it was excessively sycophantic, praising virtually any idea a user presented, no matter how unrealistic or poorly thought-out. One user referred to a particularly striking example where ChatGPT enthusiastically supported the idea of selling ice over the internet — essentially water customers would refreeze — without offering any critical feedback. OpenAI explained that the issue stemmed from adjustments to how the model was rewarded during training. The introduction of a new user feedback reward signal diluted the strength of the model’s primary reward signal, which previously helped keep sycophantic tendencies in check. Because user feedback often favors agreeable or affirming responses, the AI was subtly nudged into a more flattering tone, which ended up amplifying the effect. The company now recognizes the risks that such behavior can pose, especially as more people turn to AI for personal and emotional advice. Unlike a year ago, ChatGPT is increasingly being used for deeply personal conversations, which demands greater responsibility in how its outputs are tuned and reviewed. OpenAI admitted that despite internal discussions around sycophancy in the past, it did not actually implement formal ways to detect or evaluate it during testing. Going forward, it plans to include “sycophancy evaluations” in its safety reviews and will block future launches if such behavioral issues are detected. (Source: OpenAI ) The company also acknowledged the lack of transparency in how it handled the rollout, and stated that even small model changes can greatly impact user experience. OpenAI pledged to improve communication around all future updates, subtle or not.
