"Change the world" is a lofty goal. If you're just starting out, I'd suggest changing your community.
Start with an under-represented population (an ethnic group, gender, or class) in your geography. Then, identify a problem affecting/limiting the ability of an individual in that group to change their circumstances (literacy, access to finance, etc.). When you figure out whether you can code a solution, it is probably a solution that is scalable to other communities.
If you can't code the solution directly, examine other ways you can automate assisting them. For example, letter writing is still regarded as an important means of reaching out to US politicians to address an issue. That means, if you are illiterate or of limited literacy, you're probably unable to have a voice in US politics.
Of course, to do all this, you need to find the population you want to serve and reach out to them. That's why I suggest starting in your community, rather than starting with "the world".
Look me up if you'd like to discuss this further – I used to work in refugee aid in Istanbul and there were many ways a developer could have helped us fight human trafficking, identify refugees at risk of deportation, or just ensure refugees were processing their paperwork correctly. I'd be happy to share my experience.