Thought process:

Overarching problem: Need to find security vulnerabilities you can solve on a regular basis or with a service, so you need to find common attacks for which people are willing to pay for protection/resolution

It's better to be a saas than a contractor (stability wise). We are concerned with anything that threatens any kind of security. What is the most dangerous thing in the world currently? We want something that integrates AI into the solution in a meaningfull way. I believe it may be advantageous to think of the word security in a more loose manner. Here are areas I feel security is lacking:

Job security
Financial stability
Information security
Identity/reputation security

...

Anyways, we settled for identity/reputation security as it was the option that seemed the most promisinig for AI integration as there seemed to be a gap in the market; there are multiple well known companies in the sector such as incogni, deleteme and aura to name a few, but none of them provide the full suite of services that would allow you to protect yourself in one convenient package. As it stands, it would cost arouned 270$ (minimum) for the average person to get a basic coverage from all theses services. We consider it quite feasible to offer the full extent of their services for 10-15$ if we seek to generate revenue from user data and use our own algorithm to find misusses of their likenesses.

Core idea:

We plan to collect data on people on the internet from the USA, Canada and other countries where user data is worth more money once a day, then categorize the data in a hashmap (using AI) for easy access. We can then compare data against user data the user gives us access to and find all elements pertaining to them in the hashmap and evaluate how likely a given element is to be illegitimate. For example, an add with their face for companies we have no data linking to them would be ranked closer to a one and flagged for review, whereas a picture of them at a hackathon is likely not illegitimate use and would receive a value close to 0.