As we recently reported, IBM has pledged $5 million in prizes to anyone who can turn artificial intelligence into tools that could solve some of the biggest problems in the world, like climate change or world hunger — certainly ambitious projects. But as the push for AI intensifies, many are warning that it could be more dangerous than nuclear weapons if not developed carefully.

The competition would involve independent developers building apps using the Watson AI system — the same computer famous for its Jeopardy appearances — and the competition would seek to yield new breakthroughs by 2020, according to a ComputerWorld report. The contests were announced at a TED conference in Vancouver recently, thanks to a partnership between X Prize and IBM. The winners would be announced in 2020, with three teams splitting $4.5 million and the remaining $500,000 going to other projects.

But this push for the rapid development of AI has some concerned, including Professor Stuart Russell, a computer scientist at the University of California in Berkeley and the Centre for the Study of Existential Risk at Cambridge University who is known for his work on AI, according to a Daily Mail report. He believes that mankind may be “driving off a cliff” by doing this.

The problem with AI is that it could be easily adapted for military use, and thus sentient computers could be in charge of some terrifying weaponry. Too much trust in AI could lead to problems down the road, Russell warned.

And he’s not the only one. SpaceX and Tesla founder Elon Musk and famed physicist Professor Stephen Hawking have also spoken out about the issue.

And the movement for AI began in a similar fashion as the push for nuclear technology, Russell noted. The primary interest in nuclear technology at first was its ability to supply an enormous amount of energy. Soon, however, mankind used it for weapons — which is why it’s important for mankind to be vigilant when it comes to AI.