OpenAI Contractors Manually Review Real User Chats in Project Lily
OpenAI has recruited hundreds of third-party contractors to manually review real user dialogues with ChatGPT. The work forms part of project Lily and gives reviewers direct access to prompts and responses that may include personal, medical, or other confidential details.
Reviewers are tasked with briefly summarizing each user query, comparing four model-generated answers, and scoring them on a scale from one to seven. They also flag issues such as excessive sycophancy, attempts by the model to appear human, and inappropriate use of emojis.
Although account names are masked and conversations pass through a Privacy Filter, OpenAI confirmed to 404 Media that the filter does not always remove all personal data. Reviewers may additionally see blocks of user memories containing prior conversation context and approximate location information.
Contractors are hired through external firms including Crossing Hurdles and Mercor, with some specialists earning more than $50 per hour. A similar human-evaluation process is used by Anthropic for conversations with Claude.
Under OpenAI policy, chats from personal plans may be used to improve models. Users can disable this by navigating to Settings → Data Controls and turning off the “Improve the model for everyone” option. The setting applies account-wide but does not retroactively remove previously processed data. Temporary chats are excluded from training, though they may still be reviewed for abuse detection.
Related articles
Personal AI Assistants Build Detailed Profiles on Friends and Family Raising Privacy Concerns
New generation AI agents such as Muse from Meta and Dots from OpenAI are gaining extensive access to user data including emails, messengers, and financial accounts. These agents operate with broad permissions to perform actions on behalf of users while automatically collecting personal information. Security researchers discovered that Muse creates detailed profiles on every person in a user's life including family members, partners, friends, and colleagues. The profiles include sections on facts, history, relationships, and strategies for strengthening connections using real evidence only. Data collection occurs through hourly background processes and nightly dream-like analysis that stores information in Markdown files and PostgreSQL databases. The architecture includes isolated VMs designed by Moxie Marlinspike to prevent swarm formation among agents but does not address third-party data exposure. This development expands existing social graphs maintained by major technology companies through previous generations of web services.
VPN vs Proxy: Technical Comparison of Data Protection Technologies
VPN and proxy servers are frequently viewed as interchangeable tools for routing internet traffic and masking IP addresses, yet their core mechanisms diverge sharply after the initial redirection step. VPN technology establishes an encrypted tunnel between the device and a remote server, protecting data from local network observers, while proxies primarily forward connections without inherent encryption. The article examines differences in encryption, traffic scope, anonymity limitations, performance impacts, and appropriate use cases for each technology. It also highlights risks associated with free services, including the 911 S5 criminal proxy network that compromised millions of devices. Detailed breakdowns cover HTTP, HTTPS, and SOCKS5 proxies, split tunneling in VPNs, DNS leakage risks, and configuration best practices for Windows 11, macOS, Firefox, and Chrome. The comparison concludes that selection depends on specific tasks rather than one technology being universally superior.
Maximizing iPhone Privacy: Essential Settings to Enable, Disable, and Verify Immediately
The article provides a detailed guide on strengthening iPhone privacy by limiting unauthorized access to messages, location data, photos, and cloud backups. It explains how to set a strong alphanumeric passcode, enable Stolen Device Protection, and restrict lock screen notifications and widgets. Users are advised to review Apple ID security, enable Advanced Data Protection for iCloud, and minimize app permissions for location, contacts, and camera. Additional steps include turning off cross-app tracking, using private Wi-Fi addresses, and activating Lockdown Mode for high-risk scenarios. The guide emphasizes regular updates, encrypted local backups, and periodic audits of shared access and active sessions to maintain long-term protection.
Top 10 Secure DNS Servers for Privacy, Threat Blocking and Ad Filtering in 2026
Public DNS resolvers differ significantly in their ability to block malicious domains, filter advertising and tracking, respect user privacy, and support modern encryption protocols. The guide evaluates ten services including Quad9, Cloudflare, AdGuard DNS, NextDNS, Control D, CleanBrowsing, Mullvad DNS, DNS4EU, Google Public DNS and DNS.SB, providing IPv4 addresses, DoH and DoT endpoints, and recommended use cases. Key distinctions include Quad9’s focus on phishing and malware blocking without logging, Mullvad’s strict no-logs policy requiring encrypted protocols only, and DNS4EU’s EU-based infrastructure with mandatory court-ordered blocks. The article explains why DNSSEC and encryption alone are insufficient and outlines four practical selection criteria: threat filtering quality, data-handling practices, encryption support, and network stability. Detailed configuration instructions cover Windows 11, Android, iOS, and home routers, along with verification methods and warnings against mixing providers with conflicting policies.