[PODCAST] You're Deploying AI Wrong and Here's How to Fix It w/ Rowan Trollope and Jonathan Rosenberg Pt. 3
Speech recognition, natural language processing, and automation. That’s what we need to emphasize in the future of AI. NOT self-driving cars, which is nowhere near as important as, you know, saving lives.
Take away the cost of inaccuracy.
“People drive cars and die. So let’s put software in charge of that!”
No, just no.
Instead, let’s leverage AI in diagnostics, where literal lives can be saved with accurate analysis of test results. Or even in customer service, where the worst thing that can happen is someone gets slightly upset.
Move up the stack in terms of value.
An AI that can identify whether people who are calling about mortgages live in a flood zone. Now the agent can offer flood insurance.
Or an AI that can process the speech of calls to eliminate post-call surveys. Now you already know it was a five-star call.
Assume the AI won’t work.
If we assume that AI will mess up even 30% of the time, we won’t have losses when that comes true.
Put AI in charge of things that are zero cost. When it does something great, it’s 100% gain.
Seek applications for AI where you can tolerate or ignore mistakes and still get benefit when it's right.
To hear more about the future of AI, listen to the rest of our conversation with Jonathan and Rowan here.