conversational interface

By Futurist Thomas Frey

The Seductive Logic of Having a Conversation with Abstract Concepts

Can you talk to conspiracy theories? To economic recessions? To your own lack of motivation? What about contrails, unsolved crimes, the magnetosphere, or your personal biases?

The short answer: yes, if you’re willing to accept that you’re not actually talking to these things—you’re talking to AI models sophisticated enough to simulate their behavior, explain their mechanisms, and respond as if they were entities with agency.

The longer answer is more unsettling: we’re going to do this whether it’s philosophically coherent or not, because conversational interfaces are irresistibly compelling. And the results will range from genuinely helpful to dangerously misleading depending on what we’re trying to give voice to and why.

The concept of “talking to the defect”—using AI to transform complex systems into conversational partners—extends far beyond medicine. Any phenomenon that can be modeled can theoretically be given voice. But there’s a dangerous assumption embedded in this entire framework: that giving something a voice makes it more trustworthy, more comprehensible, more real.

Let me walk you through where conversational systems become transformative, where they become actively dangerous, and why we’re building the oracles first and planning to figure out the difference later.

AI & Big Data

One can intuitively surmise artificial intelligence (AI) is today’s hot commodity, gaining traction in businesses, academia and government in recent years. Now, there is data — all in one place — that documents growth across many indicators, including startups, venture capital, job openings and academic programs. These bellwethers were captured in the AI Index, produced under the auspices of was conceived within Stanford University’s Human-Centered AI Institute and the One Hundred Year Study on AI (AI100).

One key measure of AI development is startups and venture capital funding. From January 2015 to January 2018, active AI startups increased 2.1x, while all active startups increased 1.3x, the report states. “For the most part, growth in all active startups has remained relatively steady, while the number of AI startups has seen exponential growth,” the report’s authors add. The trickle of venture capital into AI startups, another bellwether, also turned into a torrent. VC funding for AI startups in the US increased 4.5x from 2013 to 2017. Meanwhile, VC funding for all active startups increased 2.08x.

Another key measure, job openings, accelerated in AI. While machine learning is the largest skill cited as a requirement, deep learning is growing at the fastest rate — from 2015 to 2017 the number of job openings requiring deep learning increased 35x, the report’s authors state.

Continue reading… “How fast is artificial intelligence growing? Look at the key bellwethers”