Subscribe via RSS Feed Connect on Google Plus Connect on Flickr

The AI Apocalypse: Warning No. 702

Monday, 17 July, 2017

Elon Musk has said it before and now he’s saying it again. We need to wise up to the dangers of artificial intelligence (AI). Speaking at the National Governors Association meeting in Rhode Island at the weekend, the CEO of Tesla and SpaceX said that AI will threaten all human jobs and could even start a war.

“It is the biggest risk that we face as a civilization,” he said.

Musk helped create OpenAI, a non-profit group dedicated to the safe development of the technology and he’s now urging that a regulatory agency be formed that will monitor AI developments and then put regulations in place. Fans of AI say such concerns are hasty, given its evolving state.

Note: Open AI and Google’s DeepMind released three papers last week — “Producing flexible behaviours in simulated environments” — highlighting an experimental machine learning system that used human teamwork to help an AI decide the best way to learn a new task. For one experiment, humans provided feedback to help a simulated robot learn how to do a backflip. The human input resulted in a successful backflip with under an hour of feedback, compared to the two hours of coding time an OpenAI researcher needed which, by the way, produced an inferior backflip to the human-trained one.

Is this important? Yes, because evidence is emerging that an AI can do some tasks better with human instruction — from cleaning someone’s home to learning a patient’s unique care needs. OpenAI hopes that if we can “train” AI to work closely with humans, we’ll be able to moderate some of the potential downsides of the technology. Like replacing journalists or starting a war.


Filed in: AI • Tags: , , , ,

Comments are closed.