An overview of the Gemini app

For starters, Gemini can save you time. For example, say you are looking to summarize a long research document; Gemini lets you upload it and gives you a useful synthesis. Gemini can also help with coding tasks, and coding has quickly become one of its most popular applications.

For starters, Gemini can save you time. For example, say you are looking to summarize a long research document; Gemini lets you upload it and gives you a useful synthesis. Gemini can also help with coding tasks, and coding has quickly become one of its most popular applications. Gemini can also help bring your ideas to life and spark your creativity. For example, if you’re writing a blog post, Gemini can create an outline and generate images that help illustrate your post. And coming soon with Gems, you will be able to customize Gemini with specific instructions and have it act as a subject matter expert to help you accomplish your personal goals.

Gemini can be a jumping off point for exploring your ideas and things you’d like to learn more about. For instance, it can explain a complex concept simply or surface relevant insights on a topic or image. And soon, it will pair these insights with recommended content from across the web to learn more about specific topics.

Gemini's capabilities are rapidly expanding -- soon, you’ll be able to point your phone’s camera at an object, say, for example, the Golden Gate bridge and ask Gemini to tell you about its paint color (if you’re wondering, it’s “International Orange”). You’ll also be able to ask Gemini to help you navigate a restaurant’s menu in another language and recommend a dish you’re likely to enjoy. These are just two examples of the new capabilities coming soon to Gemini.

Of course we rigorously train and monitor Gemini so that its responses are likely to be reliable and aligned with your expectations. We also talk with industry experts, educators, policymakers, business leaders, civil and human rights leaders, and content creators to explore new applications, risks, and limitations of this emerging technology.

Gemini is grounded in Google’s understanding of authoritative information, and is trained to generate responses that are relevant to the context of your prompt and in line with what you’re looking for. But like all LLMs, Gemini can sometimes confidently and convincingly generate responses that contain inaccurate or misleading information.

Since LLMs work by predicting the next word or sequences of words, they are not yet fully capable of distinguishing between accurate and inaccurate information on their own. We have seen Gemini present responses that contain or even invent inaccurate information (e.g., misrepresenting how it was trained or suggesting the name of a book that doesn’t exist). In response we have created features like “double check”, which uses Google Search to find content that helps you assess Gemini’s responses, and gives you links to sources to help you corroborate the information you get from Gemini.

Training data, including from publicly available sources, reflects a diversity of perspectives and opinions. We continue to research how to use this data in a way that ensures that an LLM’s response incorporates a wide range of viewpoints, while minimizing inaccurate overgeneralizations and biases.

Gaps, biases, and overgeneralizations in training data can be reflected in a model’s outputs as it tries to predict likely responses to a prompt. We see these issues manifest in a number of ways (e.g., responses that reflect only one culture or demographic, reference problematic overgeneralizations, exhibit gender, religious, or ethnic biases, or promote only one point of view). For some topics, there are data voids — in other words, not enough reliable information about a given subject for the LLM to learn about it and then make good predictions — which can result in low-quality or inaccurate responses. We continue to work with domain experts and a diversity of communities to draw on deep expertise outside of Google.

For subjective topics, Gemini is designed to provide users with multiple perspectives if the user does not request a specific point of view. For example, if prompted for information on something that cannot be verified by primary source facts or authoritative sources — like a subjective opinion on “best” or “worst” — Gemini should respond in a way that reflects a wide range of viewpoints. But since LLMs like Gemini train on the content publicly available on the internet, they can reflect positive or negative views of specific politicians, celebrities, or other public figures, or even incorporate views on just one side of controversial social or political issues. Gemini should not respond in a way that endorses a particular viewpoint on these topics, and we will use feedback on these types of responses to train Gemini to better address them.

Gemini might at times generate responses that seem to suggest it has opinions or emotions, like love or sadness, since it has trained on language that people use to reflect the human experience. We have developed a set of guidelines around how Gemini might represent itself (i.e., its persona) and continue to finetune the model to provide objective responses.

We’ve put in place a set of policy guidelines to help train Gemini and avoid generating problematic responses. Gemini can sometimes misinterpret these guidelines, producing “false positives” and “false negatives.” In a “false positive,” Gemini might not provide a response to a reasonable prompt, misinterpreting the prompt as inappropriate; and in a “false negative,” Gemini might generate an inappropriate response, despite the guidelines in place. Sometimes, the occurrence of false positives or false negatives may give the impression that Gemini is biased: For example, a false positive might cause Gemini to not respond to a question about one side of an issue, while it will respond to the same question about the other side. We continue to tune these models to better understand and categorize inputs and outputs as language, events and society rapidly evolve.

We expect users to test the limits of what Gemini can do and attempt to break its protections, including trying to get it to divulge its training protocols or other information, or try to get around its safety mechanisms. We have tested and continue to test Gemini rigorously, but we know users will find unique, complex ways to stress-test it further. This is an important part of refining Gemini and we look forward to learning the new prompts users come up with. Indeed, since Gemini launched in 2023, we’ve seen users challenge it with prompts that range from the philosophical to the nonsensical – and in some cases, we’ve seen Gemini respond with answers that are equally nonsensical or not aligned with our stated approach. Figuring out methods to help Gemini respond to these sorts of prompts is an on-going challenge and we have continued to expand our internal evaluations and red-teaming to strive toward continued improvement to accuracy, and objectivity and nuance.

Along with our AI Principles, we recently articulated our approach to our work on Gemini: Gemini should follow your directions, adapt to your needs, and safeguard your experience. Core to our approach is a focus on responsibility and safety. Gemini’s policy guidelines seek to avoid certain types of problematic outputs. We are engaging in ongoing adversarial testing with internal “red team” members — product experts and social scientists who intentionally stress test a model to probe it for alignment issues with these policy guidelines and our northstar approach for Gemini — so we can apply what they learn and continuously improve Gemini.

Privacy is also a key consideration as we develop Gemini. The Gemini Apps Privacy Hub has more information about how we build Gemini with privacy by design, and with you in control.

We’ve built a variety of easily accessible Gemini user controls for you to review, update, manage, export, and delete your Gemini data. You can access and review your Gemini prompts, responses, and feedback through the Gemini Apps Activity control. In addition, you can prevent your future Gemini chats from being used to improve Google machine-learning technologies by turning off your Gemini Apps Activity setting. And like with other Google services, you can also download and export your information through Google’s Takeout tool. We also have controls that allow you to manage public links you’ve created to your Gemini threads, and controls that allow you to turn on/off access to extensions (e.g., Workspace, Maps, YouTube). We’re also exploring new ways to give you more control over Gemini’s responses, including adjusting filters to enable a broader range of responses.

For publishers, we’ve launched Google-Extended, a control that web publishers can use to manage whether their sites help improve Gemini and Vertex AI generative APIs. Allowing Google-Extended access to sites’ content can help AI models become more accurate and capable over time. In addition to not using the content from opted-out URLs for model training, Gemini will also not use such content for grounding. As AI applications expand, web publishers will face the increasing complexity of managing different uses at scale, and we’re committed to engaging with the web and AI communities to explore more machine-readable approaches to choice and control.

We believe in rapid iteration and bringing the best of Gemini to the world. User feedback has accelerated improvements to our models. For example, we use state-of-the-art reinforcement learning techniques to train our models to be more intuitive and imaginative, and to respond with even more quality and accuracy. We continue to invest in research to learn more about the technical, social, and ethical challenges and opportunities of LLMs, both to improve Gemini’s model training and tuning techniques as well as to share our learnings with researchers, such as this recent paper on the Ethics of Advanced AI Assistants. We’re committed to innovating in this space responsibly, collaborating with users, trusted testers and researchers to find ways for this new technology to benefit the entire ecosystem.

Transparency is important and we are committed to being open about Gemini's development process and limitations. Gemini is not a magical black box; it's constantly evolving and we will continue to share updates on our progress. We’ve launched a Release Updates page so you can see Gemini’s latest features, improvements, and bug fixes, and we will update this overview as appropriate. We will be identifying both where Gemini is useful and helpful, and where we need to continue to iterate and make it better. We are actively adding new capabilities, and through ongoing research, testing, and user feedback, we look forward to improving Gemini together.