The Unseen Risks of Gen AI: Why Businesses Hesitate

The Unseen Risks of Gen AI: Why Businesses Hesitate

Introduction

In this intriguing episode of Saturday Architecture, Kumaran and Deepak from Microsoft delve into the complexities of adopting Generative AI (Gen AI) in business. They discuss the perceived risks, costs, and challenges that come with integrating Gen AI into mainstream business operations.

???????????? ?????? ??????????????????

Kumaran:

  • Risks of Gen AI: Too risky and expensive without guaranteed outcomes.
  • Uncertainty in ROI: ??Comparing Gen AI to unpredictable stock market investments.
  • Hesitation in Adoption: ?? Lack of traction from demo to production due to reliability concerns.

Deepak:

  • Hype Flattening: ?? Gen AI hype is dying down; now it’s about real value.
  • Usability Challenges: Not always intuitive, especially for non-tech-savvy users.
  • Productivity vs. Automation: ?? Gen AI tools like Copilot enhance productivity but don't replace human tasks.

?????????????? ????????????????

Kumaran:

  • Real-world Example: Challenges of introducing Copilot in Excel to finance and operations departments.
  • Cognitive Load: ?? Too taxing for users to figure out if and when it works.
  • Adoption Strategies: Need for clear, reliable use cases and user enablement.

Deepak:

  • Adoption Similarities: Same rules apply to Gen AI as to any tech adoption.
  • Specific Use Cases: ?? Identifying scenarios where Gen AI always works and others that need user training.
  • Learning Curve: Need to approach like learning to drive a new tech-enhanced car.

?????? ?????????????? ???? ?????????? ????????????????

Kumaran:

  • Implementation Strategy: ??? CIOs need to classify use cases by reliability and train users accordingly.
  • Ensuring Predictability: ?? Combining user enablement with reliable scenarios to build confidence.

Summary & Action Items

Summary:

  • Gen AI adoption faces challenges due to perceived risks and costs.
  • Clear, reliable use cases and user training are essential for successful implementation.

Action Items:

  1. Identify Reliable Use Cases: CIOs should pinpoint scenarios where Gen AI is 100% reliable.
  2. User Training: Develop comprehensive training programs to enhance user understanding and confidence.
  3. Build Confidence: Gradually roll out Gen AI tools, starting with the most predictable scenarios.



Ramya Sampathkumar

Chief Information & Digital Officer, Brakes India | Global Top 100 CDO 2022 | Strategy to Change

4 个月

Nice discussion. I think if you break this into its bare bones, there are broadly 3 types of applications- mission critical apps where there can be no fault tolerance, apps that respond to your queries in a pre determined transactional way, and apps that are exploratory and context driven. The way you handle training, adoption and broader change management for each will differ. GenAI falls in the last category and your summarisation captures the approach well.

Bharathi S.

A learner. My Ikigai - To learn new skills/perspectives from both successes & failures and enable others to do same

5 个月

Kumaran, Thanks for interesting discussion needed in the current context. While adoption and maturity level of adoption will take time, most important focus for organisations would be increasing literacy on GenAI within the tech and non tech community. This will have a direct corelation in reducing the cognitive load on the end users, business community, practitioners.! Exciting era of learning is awaiting

Kartic Vaidyanathan

Founder @ LetUsPlayToLearn | Social Networking, Coaching & Mentoring | Guest Faculty, IIT Madras, PPD

5 个月

Shree Krishna Priya J - Thought this blog discussion might be relevant to what you are trying in your post series where you are attempting to prepare yourselves and others for AI replacing humans and how they need to prep themselves in some form. Not in a very direct sense but there is still a lot of uncertainities and doubts in adoption that would have to be addressed before it gets adopted. Some pointers discussed by Kumaran and Deepak might be of use.

要查看或添加评论,请登录

Kumaran Anandan (????? ??)的更多文章

社区洞察

其他会员也浏览了