ZOOM SUMMIT
The path to responsible AI
Takeaways from Zoom’s Transformation Summit
Earlier in the year, Zoom convened its 2023 World Transformation Summit, bringing together business executives, artificial intelligence (AI) practitioners and thought leaders across industries. The goal: share strategies for leveraging AI to enhance customer and employee experiences responsibly. Spanning keynotes, fireside chats and panel discussions, the virtual event unpacked best practices for implementing AI amid mounting hype. Sessions also covered risks including algorithmic bias and misinformation along with mitigation tactics. Here are the summit’s most actionable insights for charting an ethical path to AI adoption.
Infuse AI into Work Thoughtfully Keynote speaker Pascal Bornet expanded on judiciously embedding AI into business operations. As an AI practitioner, Bornet focuses on “making the world more human using technology.” To Bornet, AI’s value is freeing employees from repetitive, low-value work so they can undertake higher-level analysis and decision- making. However, he stressed that, unlike human colleagues, AI lacks contextual understanding. Treat it as a collaborative digital asset, not an independent worker. Bornet urged organisations to identify opportunities where AI can absorb tedious, rules-based tasks while augmenting complex cognitive processes. Form cross-functional teams to pinpoint use cases, calculate potential efficiency gains and continually evaluate AI’s impact on overall productivity and output quality. Deliver Personalised, Emotionally Intelligent CX According to Metrigy CEO Robin Gareiss, natural language processing and sentiment analysis make conversational AI invaluable for boosting customer satisfaction. Beyond fielding common support questions, AI chatbots can explain bill charges, recommend
Smita Hashim Chief Product Officer
Curb AI Enthusiasm with Human Oversight
www.zoom.us
In her opening keynote, Zoom Chief Product Officer Smita Hashim acknowledged feeling awestruck by AI’s rapid evolution. The ability to have natural conversations with chatbots like Claude and GPT-3 feels magical. However, Hashim cautioned the AI hype warrants perspective. Generative language models remain computational tools devoid of human judgement and context. Despite their eloquence, they frequently generate convincing but inaccurate or nonsensical information. To benefit from AI’s capabilities while controlling for quality, Hashim advocated a “human-centred design” approach. For instance, enable users to review and edit AI-generated meeting summaries before disseminating them. Let humans guide the technology. Hashim also discussed Zoom’s “federated approach” to AI, leveraging a mix of proprietary models, third-party services like Anthropic and customer-specific customisation based on unique vocabulary and use cases. This allows Zoom to advance with the AI field instead of locking into any single provider.
Keynote
products, alert customers to status changes, and channel complex issues to human agents. Gareiss noted AI’s ability to interpret customer emotions
Pascal Bornet
CONTINUED
44
Powered by FlippingBook