Generative AI is not the panacea we’ve been promised | Eric Siegel for Big Think+

Summary

Eric Siegel, co-founder and CEO of Goodr AI, addresses the hype around generative AI, emphasizing its impressive but limited capabilities compared to humans. He distinguishes it from predictive AI, which offers untapped value in improving large-scale operations through data-driven predictions. Siegel illustrates predictive AI's practical applications in industries like delivery, highlighting UPS's efficiency gains and cost savings. The key takeaway is to focus on concrete value and specific use cases rather than being swayed by hype.

Takeaways

• 🧩 Generative AI is often perceived as a panacea capable of solving all business problems, but this is an overstatement and a form of hype.
• 🤖 Eric Siegel, CEO of Goodr AI, emphasizes that while generative AI is impressive, it is not going to run the world and is more limited in its capabilities compared to predictive AI.
• 📝 Generative AI can create first drafts for documents but requires proofreading and cannot be trusted blindly, indicating a need for human oversight.
• 🔮 Predictive AI, which is older, still has untapped value and is more suited for large-scale operations, unlike generative AI which is often limited to per-word level understanding.
• 🛠 Predictive AI is used for improving existing operations by learning from data to make predictions that can automate and prioritize decisions in various industries.
• 🚚 UPS uses predictive AI to streamline delivery efficiency, saving millions of dollars annually and reducing emissions, demonstrating the practical application of predictive AI.
• 📊 Predictive AI involves working with probabilities to make informed decisions, which is crucial for improving large-scale operations.
• 🔑 The value of AI comes from its deployment and the changes it brings to existing operations, rather than the technology itself.
• 🧐 Siegel expresses skepticism about the hype surrounding the potential of AI to achieve AGI (Artificial General Intelligence), cautioning against overestimating its capabilities.
• 🌟 The seemingly human-like capabilities of generative AI are fascinating, but Siegel advises focusing on concrete value and practical applications rather than philosophical debates.
• 📈 To effectively utilize AI, businesses should identify specific, credible use cases that can deliver tangible improvements to enterprise operations.

Q & A

What is the main illusion associated with generative AI according to the transcript?

The main illusion is that generative AI is on the brink of solving all business problems automatically and is a potential panacea, which is actually hyperbole and hype.

What is Eric Siegel's professional background?

Eric Siegel is the co-founder and CEO of Goodr AI, the founder of the Machine Learning Week conference series, and the author of 'The AI Playbook: Mastering the Rare Art of Machine Learning Deployment.' He has been in the field of machine learning since 1991.

How does Siegel describe the capabilities of generative AI like chatGPT?

Siegel describes generative AI as capable of communicating about any topic and often giving responses that seem to understand what you're saying, but he emphasizes that its understanding is limited to a low level of detail, the per-word level.

What is the difference Siegel sees between the capabilities of generative AI and human understanding?

Siegel believes that the difference will become increasingly apparent as generative AI operates on a low level of detail and often gets things right as a side effect, whereas human understanding is much deeper and nuanced.

What is the potential use of generative AI in writing according to Siegel?

Generative AI is valuable for writing first drafts, such as letters or syllabi, but it cannot be trusted blindly and requires proofreading.

What is predictive AI and how does it differ from generative AI?

Predictive AI is technology that learns from data to make predictions in order to improve large-scale enterprise operations. It differs from generative AI in that it focuses on improving existing operations through predictions rather than creating content.

How does Siegel view the potential of predictive AI in improving large-scale operations?

Siegel views predictive AI as having great untapped value and potential for autonomy, as it can systematically make decisions over and over again, fully autonomously, leading to significant improvements in operations.

Can you provide an example of how predictive AI is used in the real world as mentioned in the script?

An example given is UPS using predictive AI to streamline the efficiency of their deliveries by predicting tomorrow's deliveries, which results in annual savings of three hundred and fifty million dollars and reduction in emissions.

What is the importance of acting on the predictions made by predictive AI according to Siegel?

Acting on the predictions is crucial because the value of predictive AI comes from deploying it and changing existing operations, not just from the accuracy of the predictions themselves.

What is Siegel's stance on the hype surrounding Artificial General Intelligence (AGI)?

Siegel is skeptical about the hype surrounding AGI, stating that he does not believe we are close to fully replicating human capabilities with computers and that such expectations are mismanaged.

What advice does Siegel give to counter the hype around AI technologies?

Siegel advises to focus on concrete value, determine specific, credible use cases for AI technologies, and be practical rather than overly optimistic about their capabilities.