Special issue: predictions

What AI still can’t do

Artificial intelligence won’t be very smart if computers don’t grasp cause and effect. That’s something even humans have trouble with.

Humans are responsible for 25% to 40% more of the total share of methane emissions than previously estimated, according to a new study in Nature....

The methane problem: Methane is one of the most potent greenhouse gases: about 28 times more effective than carbon dioxide at trapping heat in the atmosphere, it’s responsible for about a quarter of global warming. It’s produced naturally by animals, volcanoes, and wetlands, but it’s also a byproduct of oil and gas production. It’s this last form of methane that the study focused on.

How they worked it out: Researchers used ice core measurements from Greenland from 1750 to 2013, plus previous data from Antarctica. They melted the ice to release the small quantities of ancient air trapped inside. These act a bit like time capsules, allowing us to get a snapshot of the methane in the atmosphere at the time. They used the isotope carbon-14, which comes from living things, as a proxy to determine whether the methane they found came from biological sources. Until 1870, around the time we started using fossil fuels, almost all methane came from these sources. After that, there was a rise in methane that didn’t have any carbon-14, from ancient fossil sources in which the isotope had disappeared. That allowed the researchers to compare natural methane with methane caused by human activity.

A possible upside? If more methane is created by humans, there’s an even bigger opportunity to rein in how much we release. Methane stays in the atmosphere for only a decade (compared with 200 years for carbon dioxide). So efforts to cut methane, which mostly comes from the production and transportation of gas and oil, could pay big dividends right away.

Sign up here to our daily newsletter The Download to get your dose of the latest must-read news from the world of emerging tech. 

Expand

Regulators, rein us in: Tesla and SpaceX CEO Elon Musk has said development of advanced artificial intelligence, including AI created by his own companies, should be regulated. He tweeted the remark...

Timely: The European Union unveiled a plan today to regulate “high risk” AI systems. New draft laws are expected to follow at the end of 2020. Last year 42 different countries signed up to a promise to take steps to regulate AI. However, the US and China currently seem to be prioritizing innovation and establishing supremacy in the field of AI over regulation and safety concerns. 

Long-standing worries: This is far from the first time Musk has expressed concerns about the potential negative consequences of AI development. He’s previously described it as “our biggest existential threat” and “potentially more dangerous than nukes.” In 2018 he told Recode that he thought a government committee should spend a year or two “gaining insight about AI” and then come up with regulations to ensure that it is developed and used safely.

To have more stories like this delivered directly to your inbox, sign up for our Webby-nominated AI newsletter The Algorithm. It's free.

Expand

EmTech Digital 2020

Building AI you can trust

At MIT Technology Review's AI conference, come find ideas, inspiration, and practical guidance on deploying AI in organizations in 2020.