Technology
Wednesday, February 28th, 2024 5:10 pm EDT
Key Points
- Google CEO Sundar Pichai addresses AI mistakes:
- Gemini image-generation feature taken offline for further testing due to offensive and biased responses.
- Issues reported to have offended users and shown bias, prompting Pichai to acknowledge them as “problematic.”
- Introduction and withdrawal of Gemini image generator:
- Google introduced the image generator through its main group of AI models, Gemini, allowing users to create images by inputting prompts.
- Historical inaccuracies discovered by users within a week, leading to the feature’s withdrawal for relaunch in the future.
- Google’s response and commitment to improvement:
- Pichai emphasizes the company’s commitment to providing helpful, accurate, and unbiased information across all products, including emerging AI technologies.
- Teams working on addressing issues with a clear set of actions and structural changes planned to ensure improvements in AI systems.
- Despite setbacks, Google aims to leverage recent advancements in AI models and infrastructure to build trustworthy and beneficial products deserving of user trust.
In a recent memo, Google CEO Sundar Pichai acknowledged the company’s missteps in its artificial intelligence endeavors, particularly regarding the Gemini image-generation feature. Pichai labeled the issues as “problematic” and admitted they had offended users and exhibited bias. Initially introduced as part of Gemini, Google’s AI model suite, the image generator allows users to input prompts to create images. However, historical inaccuracies surfaced within a week of its launch, prompting Google to take the feature offline for further evaluation and testing. Pichai emphasized that such biases and offensive responses were unacceptable, recognizing the high expectations placed on Google in the AI industry’s nascent stage. The memo outlined Google’s commitment to rectifying the situation, with teams working diligently to address the issues and implement structural changes, improved launch processes, and clear action plans. Pichai reiterated Google’s mission to provide users with helpful, accurate, and unbiased information across all products, including emerging AI technologies. Despite the setback, Pichai highlighted recent advancements in AI models and emphasized the company’s dedication to creating trustworthy and beneficial products. He urged a focus on building products worthy of users’ trust, leveraging Google’s infrastructure and research expertise to navigate the challenges of the AI landscape.
For the full original article on CNBC, please click here: https://www.cnbc.com/2024/02/28/google-ceo-tells-employees-gemini-ai-blunder-unacceptable.html
