Hey, guys
We would like to address recent issues with problematic text and image responses in the Gemini app (formerly Bard). We understand that some of the answers may be offensive and show bias. Let me be clear: this is totally unacceptable and our decision was wrong.
Our team is working around the clock to address these issues. We’re already seeing significant improvements in various prompts. No AL is perfect, especially in this new phase of industry development, but we know the hurdles are high for us and we will keep trying no matter how long it takes. We will then review what happened and fix it at scale.
Our mission to organize the world’s information and make it universally accessible and usable is a sacred one. We have always sought to provide our users with useful, accurate and unbiased information about our products. That’s why people trust them. This must be our approach for all products, including emerging Al products.
We will be driving a clear set of actions, including structural changes, product guideline updates, launch process improvements, robust evaluation and red teaming, and technical recommendations. We will consider all of this and make any necessary changes.
As we learn what went wrong here, we need to build on the product and technical announcements we’ve made at Al over the past few weeks. This includes some fundamental advances in the underlying model. For example, breaking through the 1 million long context window and the open model, both of which have been well received.
We know what it takes to create great products used and loved by billions of people and businesses, and our infrastructure and research expertise give us a great starting point for the Al wave. doing. Focus on what matters most: building useful products that users can trust.
