Predictim, an app for vetting babysitters, is being investigated by Facebook and has been blocked by Twitter for violating policies by using algorithms to scour applicants’ social media activity, BBC News reported.

The furore appears to centre on the US-based service’s use of artificial intelligence (AI) agorithms to probe potential child minders’ social media accounts for behaviour it describes as “aggressive, abusive, explicit, or offensive”.

On its website, Predictim said this information reveals key personality traits of potential babysitters, with the algorithms accessing “billions of social media data points” and its proprietary technology already delivering a “very high accuracy rate”.

However, the AI-based checks go against data protection policies put in place by Facebook and Twitter.

Earlier this month Facebook revoked Predictim’s access to users and is now looking into blocking it completely, as the company claimed it was still using public data on the social media site. “Scraping people’s information on Facebook is against our terms of service. We will be investigating Predictim for violations of our terms”, a Facebook representative told BBC News.

As for Twitter, it said “we strictly prohibit the use of Twitter data and APIs for surveillance purposes, including performing background checks”. Predictim was using Twitter’s API to analyse users’ tweets.

Predictim CEO Sal Parsa defended the company’s actions, telling BBC News: “Everyone looks people up on social media…we’re just automating this process.”

The app has also recieved flak in the past, with observers stating an algorithm cannot be trusted to judge a person’s employability.