SBird’s direction is Chief Product Officer of “Responsible AI” at Microsoft and is responsible for ensuring that the products Copilot and collaboration with OpenAI are safe. The goal is to prevent AI from causing harm, treating people unfairly, or spreading false content. In an interview with Financial TimesBird points out that Generative AI is different and more exciting than previous technologies because it can communicate in human language, understanding jargon and individual expressions. This makes it a bridge to other technologies, empowering people in unprecedented ways.
Index of topics:
Toggle
The impact of AI on work and its implementation
According to Bird, AI will change the way we work, as has happened with other technologies in the past. An example is GitHub Copilot, an application that completes code automatically, making developers faster and happier. Technology eliminates repetitive tasks, allowing people to focus on what adds unique value. However, it is crucial to balance product release with safety, ensuring they are tested and ready for public use without causing harm.
Collaboration with OpenAI and the road to AGI
Sarah Bird says the partnership with OpenAI is based on shared values of responsibility and safety in AI. While OpenAI focuses on developing basic models, Microsoft works on implementing external security and monitoring systems. Bird believes that theartificial general intelligence (AGI) is not a primary focus, preferring to focus on systems that augment human intelligence. While there have been significant advances, such as the leap from GPT-3 to GPT-4, there are still key elements to resolve before achieving AGI.
Challenges and responsibilities in using AI
For Bird, Key challenges include managing AI risks and hallucinations, where the system generates responses that are not aligned with the input data. Microsoft has developed capabilities to detect and correct these discrepancies in real time. Furthermore, AI must be customized to fit individual needs, while maintaining user privacy and trust. Finally, address issues like i deepfake requires a holistic and innovative approach to ensure that AI-generated content is identifiable and safe.
Personalization and inclusiveness in AI
For the Chief Product Officer of Microsoft’s “Responsible AI”, a fundamental aspect of AI is its ability to personalize, which allows it to adapt to the specific needs of users. This is particularly important to overcome the limitations of systems designed for an average user, which may not adequately respond to individual differences. Personalization not only increases the effectiveness of AI, but also promotes inclusivity, making the technology more accessible to people with different backgrounds and needs.
Addressing bias and ensuring fairness
Bias is a persistent problem in AI, and Microsoft is committed to ensuring its systems are fair and representative, Bird says. This involves extensive testing to identify and correct any biases in the data and models. The goal is to ensure that AI does not perpetuate stereotypes or injustices, but rather contributes to fair and respectful treatment of all users.
Continuous innovation and the future of AI
Sarah Bird recognizes that innovation in AI is an ongoing process. While current technologies have made great strides, there is still much to explore and develop. Collaboration with partners like OpenAI is essential to push boundaries and meet future challenges. The ultimate goal is to create systems that not only enhance human capabilities, but do so safely and responsibly, contributing positively to society.
Who is Sarah Bird
Sarah Bird works to accelerate the adoption and positive impact of AI by bringing together the latest innovations in research, products and policy to enable the use of new AI technologies such as generative AI models.
As a responsible AI implementation expert, she helps develop and adopt responsible AI principles, best practices, and technologies across the company. Bird led the cross-company expert team to develop the new Bing responsibly and led the development of responsible AI for Github Copilot.
Bird has led the development of responsible AI tools, including Fairlearn, SmartNoise e InterpretML. He is an active member of the Microsoft committee AETHER and helped create the Microsoft Responsible AI standard. Sarah was one of the founding researchers of the Microsoft Research Group FATE.
