You don’t even need to make a purchase or visit a website for data science companies to collect information about you. There are all kinds of public data, from property tax records to company and university information, aggregated through startups such as Enigma while Thasos and Reveal Mobile sell pedestrian geolocation data.

“I can show that someone saw that ad [online] and actually went into the store,” Reveal Mobile CEO Brian Handly told the audience how his startup makes user profiles to map diverse data about individuals at the Artificial Intelligence & Data Science conference in New York City. “Then using that to enhance the advertising targeting.”

Reveal Mobile passively collects data from more than 50 million phones per month through local news apps, weather apps, travel apps such as Roadtripper and many more. Most people don’t realize that opting in to an app’s location tracking features means third parties can sometimes collect and sell personal data.

There are all kinds of scenarios where layering alternative data could be beneficial, not just profitable. For example, urban planning initiatives or public hospitals that need to understand crowding and workflow without relying on insufficient credit card records. Handly told IBT his legal team is busy preparing for the European Union’s new privacy law that starts in May 2018, the General Data Protection Regulation (GDPR). “It’s impossible to say we’re compliant because we don’t know yet,” Handly said. “We’ve spent a lot of time and effort to make sure we’re up to date with GDPR.” Only time will tell how this new regulation will play out for startups who collect, analyze and sell alternative data insights.

“I think privacy by design is incredibly important,” Pavan Arora, head of data at IBM Watson, told IBT. “I think in some ways Europe is ahead of everyone else in this with GDPR. It’s going to be a standard privacy policy around personal, identifiable data.” The GDPR also applies to American companies with any European clients or sources, which extends some of the regulations protections to Americans by default. Local efforts to reform privacy laws have been slow at best. Illinois Governor Bruce Rauner vetoed the Geolocation Privacy Protection Act in September because he argued it would jeopardize many jobs.  

Most of the time, startups anonymize personal profile data anyway, meaning it isn’t attached to the user’s real name. Most data buyers are interested in trends and anomalies, not stalking specific mobile users. But there’s still the issue of securing this big data. As the Equifax hack proved, centralized data sources are prime targets for hackers. “Reputational risk is real,” Arora said. “Don’t keep it. Just don’t. Don’t keep any personal, identifiable data...as long as you maintain privacy by design from the beginning, you’ll find you are both safe and creating value.”

Even without location tracking apps, it’s hard to avoid the explosion of geolocation data craze spreading across industries like advertising, finance and prediction markets. Cell tower triangulation can map movements, albeit with less accuracy than personal mobile data. Hardly said iOS and Android make sure data-collecting apps uphold American privacy laws. Yet Quartz reported some Android phones collect user location data for Google even if the user turned off location features and doesn’t have a SIM card. Meanwhile, the startup Orbital Insight use drone and satellite data to capture information flows beyond devices. This data can be a game changer for monitoring transit and prices in global supply chains like the oil market.

Arora is excited about the opportunities he believes will come to a head in 2018 as more financial institutions combine data science with artificial intelligence. “Data trading, right now, is primitive,” he said. “I can’t believe most trades are done by algorithms, are only leveraging 20 percent of the world’s knowledge. What happens when we open up the floodgates to the other 80 percent?”

IBM Watson’s AI technologies are working on how to incorporate sentiment analysis, natural language processing and image scanning into these big data sets, not just numbers and locations. When that happens, stock market predictions will be able to leverage AI on a whole new level. “We’re talking about months, not years,” Arora said. “This is months away. We’re in for a fun ride.”

Editor's Note: Newsweek Media Group and International Business Times partnered with Structure to host this week's Artificial Intelligence & Data Science event.