The challenge sounds daunting: Use artificial intelligence "to address humanity's Grand Challenges." But if that sounds like the kind of thing that stimulates your creative juices, and you have the engineering chops to back it up, you could win a few million bucks and a lot of attention.

IBM Watson and XPrize announced earlier this year that they'd be linking up to offer a top award of $3 million and additional prizes totaling $2 million to winning entrants in the competition.

Roughly 3,500 parties responded to the announcement of the prize in February to say they were interested in competing. The five-month registration window ends Dec. 1, after which applicants will go through four selection rounds over the course of four years.

The prize money isn't much in the scheme of things when you consider the tens of millions of dollars top-raising AI startups have drawn in Series A funding rounds. It's also important to remember that funding comes in at the end of the four year competition, if you win--so it's up to competing teams to fund their entries up to that point.

But IBM Watson VP Stephen Gold says you don't necessarily need a lot of money to come up with a winning project. There are a lot of open-source resources for machine learning and artificial intelligence applications.

Here are some tips from IBM and XPrize about how to best succeed in the competition.

Remember, it's not about the money.

As Gold mentions, the $5 million in prizes aren't enough to push a major project forward. What teams stand to win by competing is attention for their project. Some are likely to get funding from venture capitalists or angel investors, but others are likely to do well on shoestring budgets. Ultimately, he thinks teams can leverage open-source resources to produce something competitive.

Partner up.

XPrize CEO Marcus Shingles and Director of Technical Operations for the IBM Watson AI XPrize Ed McNierney say collaboration is key to succeeding in the competition. Groups that have identified problems that could be solved through application of machine learning to process data, for example, should reach out to experts in machine learning and vice versa. Another component of a balanced team might be an organization with access to data relevant to the project. A team working to improve traffic safety may want government data on accidents, for example. "We want to encourage people who are trying to solve problems jump in and we'll connect you with the AI community," says McNierney.

Results trump ideas.

Shingles says those who can demonstrate success before the competition draws to a close are likely to perform the best. It's wise to find an application for the solution a team designs.

Published on: Jun 27, 2016