Glentoran call Extraordinary General Meeting as owner Ali Shams Pour looks to strengthen grip on club
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Reuters
6 hours ago
- Reuters
Telehealth's GLP-1 boom: balancing obesity care with HIPAA and state consumer privacy laws
August 20, 2025 - Demand for GLP-1 agonists, such as semaglutide and tirzepatide, exploded in 2024 and shows no signs of slowing in 2025. It is forecasted that the U.S. market will top $30 billion by the end of 2025 aided by physical brickandmortar clinics and national telehealth startups that offer fast online consults and front door delivery. The digital channel is attractive to patients as it allows for convenience and secrecy, including no waiting room, discreet shipping. As for providers, they can scale nationally. However, that same frictionless nationwide telehealth model magnifies privacy and cybersecurity risks. Online GLP-1 programs necessarily collect sensitive health histories, biometric data (such as weight, blood glucose levels, and sleep patterns), insurance details, and payment information. With that collection most providers embed advertising pixels to fuel customer acquisition. When that data straddles HIPAAregulated and consumerapp environments, the legal landscape quickly becomes treacherous. GLP-1 telehealth services almost always involve a "covered entity" (the clinician or pharmacy) plus multiple "business associates" (video visit platforms, fulfillment pharmacies, labs, and cloud vendors). Classic HIPAA safeguards therefore apply, which include encryption in transit and at rest, unique user IDs, role-based access, audit trails, and a written business associate agreement (BAA) with each vendor. Two federal developments heighten enforcement pressure: (1) Trackingtechnology bulletin: In March 2024, HHS OCR revised its 2022 guidance "Use of Online Tracking Technologies by HIPAA Covered Entities and Business Associates" clarifying that website cookies, pixels, and similar tools may transmit protected health information (PHI) when they identify a user as a patient, even if no appointment is booked. The Texas federal court decision, Am. Hosp. Ass'n v. Becerra, ---- F. Supp. 3d ----, No. 4:23-cv-1110, 2024 WL 3075865 (N.D. Tex. June 20, 2024), which vacated parts of the bulletin, created uncertainty, but OCR has reiterated that "regulated entities are not off the hook." The decision vacated only OCR's "proscribed combination" theory, that an IP address plus a visit to an unauthenticated health webpage is IIHI/PHII, leaving uncertainty about what, if any, other public-page interactions may trigger HIPAA duties when tracking tools are used. Because the rest of OCR's bulletin remains operative, especially for authenticated portals, and OCR continues to warn that disclosures of PHI to tracking vendors can violate HIPAA, entities must navigate unclear lines on public sites even as core obligations persist. (2) GoodRx's settlement consequences: The FTC's Complaint against GoodRx sent shockwaves through the healthcare industry as it was the first time the FTC had sought to use the Health Breach Notification Rule (the HNBR) (and parallel Section 5 order) against a healthcare provider. 15 U.S.C. § 45(a)(1). The FTC's Complaint alleged that GoodRx repeatedly violated these promises by sharing sensitive user information with third-party advertising companies and platforms, such as Google, Facebook and others. Specifically, GoodRx shared user prescription medication information, personal health conditions, personal contact information, and unique advertising and persistent identifiers without providing notice to the users and without first obtaining user consent to the sharing. Worse, the FTC Complaint alleged that GoodRx's sharing this information allowed the third parties to make profit from the information and use it for their own business purposes, including by using the information to personally target the users with advertisements. The suit resulted in a $1.5 million fine and a multi‑year ban on advertising using health data demonstrating the significant legal landmines when mixing telehealth services with ad retargeting. With this, OCR has indicated that it will treat similar pixel leaks as reportable breaches under HIPAA. In addition to HIPAA, 19 states, with more on the way, have comprehensive privacy laws in effect or scheduled, and several include "consumer health data" (CHD) that reaches beyond HIPAA. Three statutes matter most to GLP‑1 telehealth ventures: •Washington My Health My Data Act (MHMDA): The MHMDA covers any information "reasonably linked" to physical or mental health, including attempts to seek care, and those parts go into effect March 31, 2024 for large entities and on June 30, 2024 for small entities. The MHMDA also bans geofencing near abortion or gender‑affirming clinics and requires a signed consumer consent for CHD sharing or sale. The law also requires a signed consumer consent for CHD sharing or sale. •California Confidentiality of Medical Information Act (CMIA): The CMIA covers "Medical Information" held by any business that offers a digital health service to manage a medical condition. As expanded by AB 2089, CMIA treats telehealth weight‑management apps as "providers of health care," triggering HIPAA‑like security, access, and disclosure rules. •Florida Digital Bill of Rights (FDBR): The FDBR covers sensitive personal data, which expressly includes biometric and genetic data, and prohibits the offshoring of specific patient data and grants Floridians deletion, correction, and opt-out rights that rival those of the California Privacy Rights Act (CPRA). The FDBR also grants Florida residents deletion, correction, and opt-out rights, similar to those contained in the California Privacy Rights Act. Both CMIA and MHMDA authorize statutory damages in the amounts of $1,000 to $25,000 per violation. Plaintiffs have already alleged that pixel deployments disclose "health conditions" (obesity) and "treatment" (semaglutide) without consent, and parallel class actions under CPRA's "unauthorized disclosure" theory have been filed in California. Although GLP‑1s are not controlled substances, individual states regulate tele‑prescribing differently. Roughly one‑third require an in‑person visit before issuing an initial prescription, with Arkansas and Alabama tightening their rules this year. The issue gets even more complicated when a prescribing clinician is licensed in one state, the patient resides in another, and the compounding pharmacy is also in another because each state's respective telehealth practice laws (and privacy statute) follow with the data, meaning providers must ensure they have robust credentialing workflows and conduct conflict‑of‑law analyses before doing business or taking patients in new jurisdictions. From a privacy perspective, the shift pushes more patient data into electronic prescribing and REMS-style safety programs (Risk Evaluation and Mitigation Strategy), escalating HIPAA exposure. In response, it is recommended that telehealth platforms: (1) Map your tech stack. Take an inventory of every system that touches patient data, including intake forms, analytics, coaching apps, pharmacy routing, SMS, and reporting and identify was is covered entity, business associate or neither. Build a data flow diagram as it you would have to provide it to a regulator. (2) Minimize. For every data pathway on marketing sites and patient portals, disable non‑essential trackers by default, and negotiate "qualified service provider" status (with HIPAA‑style contractual limits) for analytics vendors that must remain; (3) Update privacy notices. Update Notice of Privacy Practices (NPP) to reflect any new manufacturer datasharing obligations. (4) Ensure vendor contract management. Reevaluate BAAs with compounding pharmacies that pivot to 503A, where prescriptions must be individually tied to a named patient. This may include adding contract indemnities for off label promotion claims that may surface in data sharing litigation with compounding pharmacies that pivot to 503A, where prescriptions must be individually tied to a named patient. This may include adding contract indemnities for off label promotion claims that may surface in data sharing litigation. (5) Implement specific jurisdictional toggles. Because GLP-1 telehealth companies frequently market nationwide, they must implement specific jurisdictional toggles in order to ensure compliance, for instance, turning off cross-site tracking for state-specific IP addresses unless express opt-in consent is obtained. (6) Pixel governance program. Apply an enterprise tracker inventory, run staticsite scans after each code push, and memorialize pixel risk assessments in HIPAA securityrule documentation. (7) Dynamic consent flows. Deploy geolocation logic to trigger "affirmative written consent" dialogs and suppress the session until accepted. (8) Tier vendors. Classify suppliers as HIPAA business associates, state CHD processors, or ordinary service providers. Flow down statespecific clauses, e.g., no "sale" or "share" of Washington CHD, right to delete Florida data within 45 days. (9) Boardlevel metrics. Track monthly privacy Key Performance Indicators (KPIs), including number of tracker removals, PHI access exceptions, deletionrequest turnaround, alongside clinical outcomes. GLP1 telehealth programs promise significant transformative benefits for millions living with obesity, yet the data they harvest is among the most sensitive in the digital economy. HIPAA remains the foundation, but a rapidly growing patchwork of state consumer healthprivacy statutes and federal enforcement actions dictate the contours of lawful virtual care. The healthcare providers that succeed will treat privacy not as a backoffice compliance task but as a strategic differentiator to build patient trust through transparency, minimize data collection, and ensure rigorous vendor oversight. Sara H. Jodka is a regular contributing columnist on privacy and data security for Reuters Legal News and Westlaw Today.
Reuters
6 hours ago
- Reuters
Israel to place $500 million, US-funded order for Boeing aerial refuelling tankers
TEL AVIV, Aug 20 (Reuters) - Israel plans to buy two Boeing-made KC-46 military aerial refuelling tankers in a $500 million deal to be financed with U.S. military aid, the Israeli defence ministry said on Wednesday. It said it would sign the contract with the U.S. government once an Israeli ministerial committee for defence procurement grants its approval. The U.S. government oversees foreign military sales and transfers to other nations. The military already operates four Boeing-made KC-46 aerial tankers, the defence ministry said in a statement. Ministry Director General Amir Baram said in the statement that the aircraft would strengthen the military's long-range strategic capabilities, enabling it to operate farther afield with greater force and with increased scope. Israel used such aerial refuelling tankers during its 12-day air war against Iran in June. The statement said the contract would include equipping the planes with Israeli systems, which it did not specify. Washington provides close Middle East ally Israel with billions of dollars each year to purchase American weapons and equipment. "The contract's scope is estimated at approximately half a billion USD (U.S. dollars) and is funded through U.S. aid," the ministry statement said. Recently, some U.S. Republicans and Democrats have questioned whether the government should continue giving Israel military aid, citing its war in Gaza and concerns over whether taxpayer dollars might be better spent on domestic priorities.
Geeky Gadgets
11 hours ago
- Geeky Gadgets
Create Your First AI Agent Assistant Without Writing a Single Line of Code
Imagine having a personal assistant that never sleeps, learns from every interaction, and can handle tasks with precision, all without you needing to write a single line of code. Sounds futuristic? It's not. The rise of no-code platforms like NA10 has made it possible for anyone, regardless of technical expertise, to create their own AI agent. Whether it's a chatbot that answers customer queries, a research assistant that sifts through mountains of data, or a productivity tool that keeps your day on track, these intelligent systems are no longer reserved for tech giants or coding experts. The best part? You don't need to be a programmer to bring these ideas to life. With a few simple steps, you can build your first AI agent and watch it transform the way you work and interact with technology. In the video below Tina Huang guides you through the essentials of creating an AI agent from scratch using intuitive, no-code tools. You'll discover how to integrate powerful AI models like ChatGPT, set up workflows, and implement safety mechanisms to ensure your agent is both effective and ethical. Along the way, you'll learn how to optimize prompts, configure outputs, and even add memory for personalized interactions. Whether you're looking to automate repetitive tasks or design a tool that enhances creativity, this journey into AI development will open up a world of possibilities. So, what could your first AI agent do for you? Let's explore the answer together. Building No-Code AI Agents What Are AI Agents? AI agents are intelligent software systems designed to perform specific tasks or achieve defined goals on behalf of users. These agents use artificial intelligence models, tools, and workflows to deliver efficient, personalized solutions. Common examples of AI agents include: Customer Service Chatbots: Automating responses to customer inquiries and improving support efficiency. Automating responses to customer inquiries and improving support efficiency. Research Assistants: Gathering, analyzing, and summarizing information from various sources. Gathering, analyzing, and summarizing information from various sources. Personal Productivity Tools: Managing schedules, reminders, and task prioritization to enhance daily productivity. The effectiveness of an AI agent is determined by its core components, such as AI models, task execution tools, and safety mechanisms. Understanding these elements is fundamental to building a successful and reliable agent. Core Components of an AI Agent Every AI agent is built on a foundation of critical components that enable it to function effectively. Below is an overview of these essential elements: AI Models: These serve as the 'brain' of the agent, processing inputs and generating intelligent responses. Popular examples include ChatGPT, Claude, and Gemini. These serve as the 'brain' of the agent, processing inputs and generating intelligent responses. Popular examples include ChatGPT, Claude, and Gemini. Tools: APIs, databases, and calendars that allow the agent to execute tasks such as retrieving data, scheduling events, or sending notifications. APIs, databases, and calendars that allow the agent to execute tasks such as retrieving data, scheduling events, or sending notifications. Knowledge and Memory: Systems that enable the agent to retain session data and access relevant knowledge bases, making sure context-aware and personalized interactions. Systems that enable the agent to retain session data and access relevant knowledge bases, making sure context-aware and personalized interactions. Audio and Speech: Features like text-to-speech and speech-to-text that assist natural language communication with users. Features like text-to-speech and speech-to-text that assist natural language communication with users. Guardrails: Safety mechanisms designed to ensure ethical, accurate, and appropriate behavior in all interactions. Safety mechanisms designed to ensure ethical, accurate, and appropriate behavior in all interactions. Orchestration: Systems for deploying, monitoring, and refining the agent's workflows to maintain optimal performance. By combining these components, you can create an AI agent tailored to meet specific user needs while maintaining reliability and efficiency. Easily Build Your First Chatbots and AI Assistants Watch this video on YouTube. Below are more guides on AI Agents from our extensive range of articles. Building an AI Agent Using NA10 Platforms like NA10 simplify the process of creating AI agents by offering intuitive, no-code tools for workflow automation and integration. Below is a step-by-step guide to building your first AI agent: Define Workflows: Start by outlining the agent's tasks. Use triggers and input forms to collect user data and initiate workflows. Start by outlining the agent's tasks. Use triggers and input forms to collect user data and initiate workflows. Integrate AI Models: Connect AI models such as ChatGPT to process inputs and generate intelligent outputs. Connect AI models such as ChatGPT to process inputs and generate intelligent outputs. Add Tools: Incorporate APIs or other resources to enable task execution, such as retrieving information or scheduling appointments. Incorporate APIs or other resources to enable task execution, such as retrieving information or scheduling appointments. Optimize Prompts: Craft clear and precise instructions to guide the agent's behavior and ensure accurate responses. Craft clear and precise instructions to guide the agent's behavior and ensure accurate responses. Implement Memory: Set up storage systems to retain session data, allowing the agent to provide context-aware and personalized interactions. Set up storage systems to retain session data, allowing the agent to provide context-aware and personalized interactions. Configure Outputs: Design the agent to deliver results in various formats, such as text summaries, visual reports, or audio files. Design the agent to deliver results in various formats, such as text summaries, visual reports, or audio files. Automate Delivery: Establish mechanisms to deliver outputs to users via email, messaging platforms, or other communication channels. This structured approach ensures that your AI agent is both functional and user-friendly, making it a valuable tool for a wide range of applications. Making sure Safety with Guardrails Safety and reliability are critical considerations when designing an AI agent. Guardrails help prevent harmful, biased, or inappropriate outputs while maintaining ethical behavior. Key strategies for implementing guardrails include: Content Screening: Use filters to detect and block inappropriate language, misinformation, or other undesirable content. Use filters to detect and block inappropriate language, misinformation, or other undesirable content. Error Handling: Develop mechanisms to identify and resolve workflow failures, making sure the agent remains functional and dependable. Develop mechanisms to identify and resolve workflow failures, making sure the agent remains functional and dependable. Ethical Guidelines: Program the agent to adhere to ethical standards, avoiding actions that could harm users or violate privacy. By incorporating these safeguards, you can build an AI agent that users can trust and rely on for accurate and ethical assistance. Orchestration and Continuous Improvement Once your AI agent is operational, ongoing management and refinement are essential to ensure optimal performance. Here's how to orchestrate and improve your agent: Deployment: Launch the agent for real-world use, making sure seamless integration with existing workflows and systems. Launch the agent for real-world use, making sure seamless integration with existing workflows and systems. Testing: Conduct thorough testing to evaluate the agent's responses, identify potential issues, and make necessary adjustments. Conduct thorough testing to evaluate the agent's responses, identify potential issues, and make necessary adjustments. Performance Monitoring: Track key metrics such as response accuracy, user satisfaction, and task completion rates to assess the agent's effectiveness. Track key metrics such as response accuracy, user satisfaction, and task completion rates to assess the agent's effectiveness. Feedback Integration: Collect user feedback to identify areas for improvement and implement updates that enhance the agent's functionality. This iterative process ensures that your AI agent remains effective, relevant, and aligned with user needs over time. Practical Applications: Research and Learning Assistant One of the most practical and impactful uses of an AI agent is as a research and learning assistant. This type of agent can: Gather information from multiple sources, such as academic databases or online articles. Summarize complex data into concise, easy-to-understand outputs. Convert text summaries into audio files for convenient, on-the-go consumption. By integrating tools like APIs for data retrieval and text-to-speech systems, you can create a versatile assistant that enhances productivity and supports continuous learning. Expanding Functionality for Future Use As your AI agent evolves, consider incorporating advanced features to improve its functionality and user experience. Potential enhancements include: Enhanced User Interfaces: Develop visually appealing dashboards and intuitive controls for seamless interaction. Develop visually appealing dashboards and intuitive controls for seamless interaction. Broader Integrations: Add support for additional APIs, tools, and platforms to expand the agent's capabilities. Add support for additional APIs, tools, and platforms to expand the agent's capabilities. Custom Features: Introduce options such as downloadable reports, advanced analytics, or multilingual support to cater to diverse user needs. These upgrades can significantly enhance the agent's value, making it a more powerful and versatile tool for users. Bringing Your AI Agent to Life Creating an AI agent without coding is now an accessible reality, thanks to platforms like NA10. By understanding the core components, implementing safety guardrails, and using orchestration systems, you can design a reliable and effective AI assistant. Whether you aim to build a research tool, a customer service bot, or a productivity enhancer, the possibilities are vast. Take the first step today and unlock the potential of AI to transform the way you work and interact with technology. Media Credit: Tina Huang Filed Under: AI, Guides Latest Geeky Gadgets Deals Disclosure: Some of our articles include affiliate links. If you buy something through one of these links, Geeky Gadgets may earn an affiliate commission. Learn about our Disclosure Policy.
