Microsoft Copilot lead says AI chatbots must ‘learn’ to ask for help to reduce hallucination episodes: “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'”

Microsoft Copilot lead says AI chatbots must 'learn' to ask for help to reduce hallucination episodes: "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'"

What you need to know

  • Microsoft’s top executive and lead of its Gen AI division says AI chatbots need to learn how to communicate when unsure of the responses generated.
  • The executive says it will help save on cost and significantly reduce instances of hallucination-generated information misconstrued as accurate.
  • Copilot helps enhance productivity at the workplace by taking over mundane tasks as professionals focus on learning new skills and building their crafts. 

As a seasoned tech enthusiast with decades of experience under my belt, I find myself both intrigued and concerned by the latest developments in AI technology, particularly the advances made by Microsoft and OpenAI. On one hand, I am thrilled at the prospect of AI tools like Copilot boosting productivity and creativity in the workplace, potentially saving companies millions of dollars annually. As a former CTO, I can attest to the fact that mundane tasks often take up an inordinate amount of time, leaving little room for innovation and skill development.


In a nutshell, when it comes to the development of Generative (or “generated”) AI systems, there are several key challenges that need to be addressed:

The situation appears unlikely to resolve swiftly, as Microsoft Vice President Vik Singh suggested during the AFP conversation.

Here’s a suggestion for rephrasing your statement in a more conversational, friendly way:

In a world where artificial intelligence (AI) tools seem to come and go, it often feels like a dance between Microsoft and OpenAI – two giants in the software and hardware industries, locked in an ongoing battle over who owns the rights to various trademarks and patents. It’s a bit like the Wild West, with both sides fighting for their own brand identity, each trying to outdo the other in terms of marketing and design. On one hand, Microsoft is focused on its Windows, Office, and Dynamics – making software, services, and apps that are designed for businesses and corporations to run smoothly; while on the other, OpenAI is all about creating content that’s meant for individuals, entertaining, engaging, and educational. It’s a high-stakes game of chess, with each side vying for the best and brightest ideas, trying to outdo the other in terms of originality, creativity, and cutting-edge technology. In this dance, Microsoft is like the big, lumbering giant that’s trying to dominate the enterprise software market – while OpenAI is like a nimble, agile startup that’s trying to shake up the consumer tech world with its flashy, disruptive innovations. It’s a classic David vs. Goliath scenario, where Microsoft is the established, traditional, and conservative force – while OpenAI is the upstart, underdog, and rebel, challenging the status quo and pushing boundaries. In a nutshell:

The CEO of OpenAI, Sam Altman, acknowledges that creating AI tools similar to ChatGPT cannot be achieved devoid of copyrighted material. He explains that while copyright laws restrict the utilization of copyrighted content, the application of such content in training AI models falls under the category of “fair use.”

Even though there’s an abundance of information available at their fingertips, AI tools sometimes veer off course and provide incorrect responses or answers to questions. As Singh explains, it is essential that these chatbots are equipped with the ability to acknowledge when they don’t have the correct answer and request assistance from a human instead.

Singh mentions that a chatbot which seeks human help 50% of the time is significantly more advantageous as it facilitates cost-saving. To clarify, each new query requires $8 to be allocated towards a customer service representative, thus substantial savings can be realized. Moreover, it enhances the customer’s experience by providing prompt responses.

Copilot helps save money and helps boost creativity

Microsoft Copilot lead says AI chatbots must 'learn' to ask for help to reduce hallucination episodes: "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'"

Recently, investors in the AI sector have expressed worries about Microsoft’s significant investments in AI initiatives, stating that it becomes challenging to identify a lucrative and straightforward business model for these projects.

Here’s a suggestion for a more conversational way of phrasing your statement:

For example, Lumen, a telecommunications company, stands to save around $50 million each year by incorporating Copilot into their daily operations. This tool streamlines repetitive tasks like research, freeing up employee time for more productive activities such as engaging directly with clients.

The major technology company intends to expand Copilot throughout its technological infrastructure, enhancing its self-reliance. However, this raises questions about employment stability within related sectors. Singh emphasizes that the technology is not intended to displace workers in the workplace. Rather, it aims to provide professionals with additional time for innovation and potentially open up new possibilities.

Read More

2024-09-03 16:41