Jump directly to the content

Smart AI divides experts as group warns it ‘would end in disaster’ as it has military links and ‘we can’t control it’

CREATING super-smart AI right now could be dangerous because we don't know enough about it to control it yet, some researchers have claimed.

According to Vox, there's a divide between AI experts working to protect us from these artificial intelligence threats.

There's reportedly a divide in the experts working to protect us from AICredit: Getty

One group is working on the current risks that AI poses.

Current dangers include the fact it can be racist, biased, and divisive.

However, there's also a group of researchers focused on the future risks of AI and how it could become too smart to control if we're not careful.

As work continues to integrate AI with military weapons, finding a way to monitor and regulate the types of AI being created is more important than ever so disaster can be avoided.

The divide in the experts working to protect us comes because some people think the AI specialists focused on the future are ignoring real current issues in favor of problems that haven't happened and may not happen.

Jack Clark, co-founder of the AI safety research lab Anthropic, recently tweeted: "Some people who work on longterm/AGI-style policy tend to ignore, minimize, or just not consider the immediate problems of AI deployment/harms."

He claimed in the same Twitter thread that "The real danger in Western AI policy isn't that AI is doing bad stuff, it's that governments are so unfathomably behind the frontier that they have no notion of _how_ to regulate, and it's unclear if they _can_".

Until we can regulate AI, some experts think it would be best to avoid making it super intelligent.

According to Vox's report, some world leaders are dismissing these concerns because they know AI has many benefits and want to speed ahead with its development.

Some experts are suggesting that countries may even need to work together to prevent the risks of AI.

Clark added: "Lots of the seemingly most robust solutions for reducing AI risk require the following things to happen: full information sharing on capabilities between US and China and full monitoring of software being run on all computers everywhere all the time. Pretty hard to do!"

Until experts can agree, teams will continue to work separately to try and monitor AI so we can reap its benefits while also staying safe.