DECEMBER 20, 2023
Google emphasizes proactive policies and adaptations to address unique risks associated with multimodal capabilities.
Gemini AI undergoes testing for biases, toxicity, and cybersecurity, addressing limitations observed in previous models like ChatGPT.
Collaboration across industries and expertise, including frameworks like SAIF, ensures a holistic approach to safety and responsible AI development.
Google works with MLCommons to develop extensive benchmarks for testing models both within Google and across the industry.
The safety and responsibility of Gemini AI are paramount considerations, with ongoing efforts to gather expert perspectives and refine the model accordingly.
Google assures users that de-identified data from free access will be used to improve the model, maintaining transparency and accountability.
As Gemini AI evolves, Google is committed to creating a secure AI framework, fostering cross-industry collaborations to enhance safety and responsibility in AI development.