Google is introducing an AI-powered Inbox view for Gmail, which aims to enhance user experience by transforming the traditional email list into a personalized interface featuring to-do items and topic summaries. This innovative approach is being rolled out initially to trusted testers in the United States and is currently available only for consumer Gmail accounts. In a demo, the AI Inbox suggests actionable tasks such as rescheduling appointments and summarizing important topics, which could help users manage their emails more effectively. However, users may face challenges if the feature generates too many to-dos, potentially leading to inbox overwhelm rather than alleviating it. Additionally, Google assures users that they can opt out of these AI features if they choose, and emphasizes that Gmail content is not used for training its AI models.
OpenAI has announced the launch of ChatGPT Health, a new feature designed to provide a secure and personalized environment for users to ask health-related questions. This product encourages users to connect their medical records and wellness apps, such as Apple Health and MyFitnessPal, to receive tailored responses. According to OpenAI, over 230 million people globally ask health questions each week, and the company has worked with more than 260 physicians to refine the model’s outputs. However, OpenAI emphasizes that ChatGPT Health is not intended for diagnosis or treatment and has faced scrutiny regarding safety, especially in handling sensitive topics like mental health. The feature operates with enhanced privacy measures, though it is not fully compliant with the Health Insurance Portability and Accountability Act. Users are advised to approach the tool cautiously, particularly those with health anxiety, as the product may inadvertently exacerbate such conditions.
Why do we care?
These tools don’t just help users do things faster—they quietly redefine what counts as work, risk, or concern.
Email has always been messy, but it was honest. Everything showed up. You decided what mattered. Google’s new inbox changes that bargain. Now the system decides what looks important, what turns into a task, and what gets summarized away.
That’s powerful—and dangerous.
Because when something goes wrong, the user doesn’t say, “I missed the email.” They say, “The inbox didn’t tell me.” And suddenly the AI isn’t just a helper; it’s the authority that failed. It’s a reimagining of the entire concept of email that’s head spinning.
Now look at ChatGPT Health. This isn’t diagnosis, but it is influence. A well-phrased explanation can delay a doctor visit just as easily as it can encourage one. When users connect wellness data, the system gains narrative authority—whether OpenAI wants that responsibility or not. And because it’s not fully regulated, there’s no clean liability boundary.
The real risk is MSPs normalizing this. Letting clients wire sensitive data into AI tools without governance. Treating AI summaries as neutral artifacts. Skipping the hard conversation about where human judgment must remain in the loop.
The business consequence is erosion of trust. Clients won’t blame “the model.” They’ll blame the environment that allowed it to speak with authority. This matters now because once AI becomes the default interpreter, walking it back is nearly impossible. The only safe move is to define boundaries before the tools define them for you.