Session Topic: Generative AI – Where are you on this journey?
Generative AI has certainly been a hot topic for IT Leaders this past year, and no doubt it will continue to be discussed by IT and business executives in the coming year. Some organizations have already adopted Generative AI tools to drive efficiency as well as innovation. Other are proceeding more cautiously, examining what guardrails and policies they need to put in place to manage potential risk.
No matter where you are on this Generative AI journey, this is a topic at the top of the minds of CIO’s as use cases for AI are explored, emerging tools are evaluated and strategies are augmented to incorporate these aspects.
Those that are already well along the path and have implemented Generative AI models, shared their use cases and learnings along the way. Those that are early in the game or just starting out on this journey, learned from their peers by listening to their stories and asking questions. No matter where you are on this journey, this was an opportunity to get involved in the discussions.
Discussion Summary:
The following questions guided the discussion and summarize the key discussion points:
What’s your company’s position use of Gen AI? If being used, do you have appropriate Governance frameworks in place?
Where GenAI is being used, it’s important to have policies and governance frameworks in place. The following are some of the points offered as advice:
- Use with guardrails, enter pilots and projects cautiously.
- Have checks in place to monitor use and outputs. Quality standards need to be defined to fact check outputs, review and validate generated content.
- Keep data private. Mask sensitive information. Don’t use company name or other identifying information in searches.
- Use for internal vs external applications. Do not expose searches externally as this exposes information to hackers.
- Own the models, and have different models federated. Information comes in pieces as models learn.
- Formulate and evolve governance as you go. Since much of AI usage is still uncharted, an approach to defining governance policies is to categorize as: light grey, grey, dark grey. Policies can be updated as usage increases and more is learned.
- AI is considered to be a Strategic Initiative with a Project Charter to formalize scope and objectives.
What are some of the exciting use cases where you have deployed Gen AI, or are in the process of deploying?
- Most are starting with Copilot as early adopters.
- Use cases are both business and IT focused.
- Start by solving small problems.
- The questions you ask are important in the quality of the generated content.
Use cases being deployed by participants include:
o Help Desk/Call Centre:
- Helping with customer communication/emails.
- Loading searches into Chatgpt to go through large volumes of data
o Chatbot:
- Finding information.
- Training.
o Unit test case generation.
o ROI analysis and Business Case building.
o Survey Results Analysis – categorizing results.
o Summarizing content from large documents or several documents on the same topic.
o Draft notes based on generated transcripts making sure to validate content.
o Process optimization.
What approaches are you taking on skilling/upskilling for Generative AI?
- Training is essential across the organization. Concepts need to move beyond IT and be beyond Chatgpt.
- Hold Lunch and Learns about concepts as well as results of pilots and usage.
- Learn as you go on a project by project basis.
How are you managing the ethics in the use of AI? Are you being transparent in the use of Generative AI?
- There needs to be defined accountability for AI implementations.
- Be authentic when using AI and disclose when AI generated content is being used.
- Business cases are generally built on efficiency. Initiatives deploying AI may not be high cost, but could be experimental or R&D, so important to watch costs.
Host: Kyoko Kobayashi
Moderators: Peter Holowka, Kin Lee-Yow, Doria Manico-Daka, Yasemin Sezer, Paul Twigg
Thanks Kyoko!