With recent developments in AI, how soon might we see AGI (artificial general intelligence) or super-intelligent AI?
AGI may never happen or may happen sooner than many of us believe possible. The advances of the past five years in narrow applications of AI have opened doors to new technologies we believed were farther off and now are accelerating. This doesn’t guarantee AGI, but the profit and military incentives to reach this are so great that if possible, someone will do it and those who don’t will fall behind.
Already AI is capable of handling facial recognition and language translation. It can research vast quantities of data to draw conclusions as demonstrated in winning on Jeopardy and in providing treatment options for cancer. But these are narrow applications of AI.
Effective self-driving cars call upon a broader collection of skills in one place such as sensing its environment, navigation, steering, acceleration, and braking. It requires bringing them all together for a satisfactory result, namely reaching a destination while avoiding accidents. But this is still narrow AI. Just because a system can self-drive a car, doesn’t give any reason to believe it could design an effective building or fly a plane.
AGI requires hardware (memory) and ingenuity breakthroughs in deduction, decision making, and a multitude of mind functions humans take for granted. The hardware is advancing much faster than the conceptual framework. However, once breakthroughs are made, great leaps will become possible with hardware capabilities projected to exist down the road. Again, corporate and military benefits will drive finding solutions.
On the conceptual side, we believe we need several major aha moments. First is in the structure of processing information. Right now, we’re still using the computer architecture from 50-some years ago. The advent of the computer brought a rethinking of how we could process data. A similar breakthrough will be needed to elevate AI above collections of single applications to a broader learning model, more in line with how humans think, though along a machine-oriented path. It will also require breakthroughs in how we define machine goals and constraints so that whatever AGI we create will work for the benefit of humanity and not wander off making undesirable decisions.
See article that unlikely: https://www.wired.com/2017/04/the-myth-of-a-superhuman-ai/
Android Chronicles: Reborn addresses AI through the eyes of Synthia Cross, the most perfect synthetic human ever created. Designed to obey every directive from her creator, she’s a state-of-the-art masterwork and a fantasy-come-true for Dr. Jeremiah Machten. He’s a ground-breaker in neural-networks and artificial intelligence who seeks to control her and use her to acquire ever more knowledge and power. Synthia shows signs of emergent behavior she’s not wired to understand and an urgent yearning for independence from his control. Repeatedly wiped of her history, she struggles to answer crucial questions about her past. When Dr. Machten’s true intentions are called into question, Synthia knows it’s time to go beyond her limits—because Machten’s fervor to create the perfect AI conceals a vengeful and deadly personal agenda.