Generative AI tools, like Microsoft's Copilot, are poised to save companies significant time and money, according to Vik Singh, a vice president at Microsoft. However, Singh acknowledges that for these AI models to be truly effective, they must learn to recognize their own limitations.
In a recent interview with AFP, Singh stated, "The thing that's really missing today is that a model doesn't raise its hands and say, 'Hey, I'm not sure, I need help.'" This shortcoming, often referred to as "hallucination," where AI generates incorrect or fabricated answers, is a critical issue for Microsoft's corporate clients, who rely on AI systems to provide accurate information without going off track.
Despite these challenges, Singh is optimistic about the potential of generative AI. He emphasized that even if AI systems need to defer to human intervention in 50% of cases, the cost savings for businesses would still be substantial. For instance, one Microsoft client spends $8 for every customer service inquiry handled by a human representative, highlighting the potential for AI to reduce operational costs and improve customer service efficiency.
Singh, who joined Microsoft in January and now leads the development of Copilot, is focused on making the AI assistant more autonomous and seamlessly integrated into Microsoft's software. The goal is for AI to handle mundane tasks, such as sending follow-up emails automatically, thereby freeing up human workers for more creative and strategic work.
This focus on productivity is central to Microsoft's strategy as it looks to justify the massive investments in generative AI. Singh pointed out that while early promises of AI "uplifting humanity" remain aspirational, the current applications are more grounded in boosting business productivity and, hopefully, profits. For example, telecom company Lumen reportedly saves around $50 million annually by using Copilot for research tasks, allowing salespeople to focus on direct customer interactions.
Despite concerns about AI potentially leading to job losses, Singh remains confident that the technology will ultimately enhance human creativity and generate new employment opportunities. Drawing on his experience at Yahoo in 2008, where AI optimized content selection for the home page, Singh noted that rather than reducing jobs, the automation led to a need for more editors to meet the increased demand for content.
As generative AI continues to evolve, Singh believes we are only in the "first inning" of what could be a transformative era for business productivity and innovation. However, the road ahead will require AI to become more self-aware and better equipped to collaborate with human workers effectively.
