Social media limit data access for AI-powered California start-up

Reading Time: 2 minutes

Predictim is an app which uses machine learning in order to vet potential babysitters. It claims to use ‘advanced artificial intelligence’ to show babysitter’s suitability and sell it to the parents. It basically combs through individual Facebook, Instagram and Twitter history to find any posts, likes or comments that could be harming for others. Founders claim, that the app can predict whether the person Is drug user, is bullying others or even have a ‘bad attitude’.

 

Experts have criticized service as unscientific and unreliable when it comes to hiring real person basing on machine learning. AI may be good in connecting facts or recognizing objects in photos but it’s not advanced enough to get jokes or sarcasm. Experts in data policy also noticed that Predictim’s software doesn’t explain how it comes to decisions, so a babysitter can lose her job without even knowing the reason.

After last week’s report by BBC News revealing that the service broke social network’s rules on data privacy and surveillance, Facebook and Twitter had taken action and instantly limited Predictim’s access to user’s data. Spokespersons of both social media said that they launched an investigation into Predictim’s extraction of personal data.

CEO and co-founder of Predctim Sal Parsa said that reaction of social medias is due to its competitive rivalry over personal data use. “Twitter and Facebook are already mining our data. It’s right there, user-generated data. Now there’s another start-up that’s trying to take advantage of that data to help parents pick a better babysitter, and make a little money in the process,” he said. “I don’t know why they ganged up on us. It could be because there’s no benefit for them.”

Although Parsa says that app is working ‘perfectly legally’, Brad Shear – specialist in social media and privacy law from Maryland, said that giving access to personal media social account by job applicants may break the law in 26 states, according to data from the National Conference of State Legislatures.

“What they are doing is purely against public policy (…). If you talk to any lawmaker out there they’ll say it’s absolutely disgusting what Predictim is trying to sell,” Shear said. “The fact they would think this is okay obviously demonstrates they have some ethical issues”.

Is AI going to substitute job interview by checking applicants’ history on social media? Is combing through personal data going to be a determinant whether person is good for a job or not? I personally think it’s going too far.
People are not reflection of their posts or comments on social medias, although they should take into consideration that everything stays in the Internet and take responsibility of what they write online. Making serious decisions only by using algorithms seem like a misunderstanding and in my opinion Predictim should make huge changes in their company’s policy.

 

References:
https://www.bbc.com/news/technology-46354276
https://www.washingtonpost.com/technology/2018/11/27/facebook-twitter-crack-down-ai-babysitter-rating-service/?noredirect=on&utm_term=.6b7107fa9284
https://www.washingtonpost.com/technology/2018/11/16/wanted-perfect-babysitter-must-pass-ai-scan-respect-attitude/?utm_term=.8b2a7d55a8cc
https://www.predictim.com/
https://www.theverge.com/2018/11/29/18117660/ai-background-checks-babysitter-predictim-data-limited-facebook-twitter-instagram

3 thoughts on “Social media limit data access for AI-powered California start-up

  1. Dzieciątko Julia says:

    According to CareerBuilder survey from 2017 70% of employers use social media to examine candidates. I have never heard about Predictim before, but there are a lot of social media and people search tools available already (Wink People Search, Hootsuite etc.). I agree that people shouldn’t be hired based only on social media content, but I wouldn’t say that the idea for Predictim itself is bad. If only they worked on legal policy…

    • Bober Joanna says:

      Yes, totally agree with you. The point is, that this app can even exclude the person from the list of candidates and applicant doesn’t have a chance to be hired. Parents can’t even take into cosideration that person, because his/her rate is so low that the app doesn’t show it as potentially good candidate. And no one is gonna search for those with bad rating. As you said, many companies are using something similar but in order to examine candidates. Unfortunately, this app doesn’t use it to do this. Instead, it research your past even before the parents can arrange job interview. So when for e.g. you write a sarcastic joke on your twitter account, you can be out… i don’t think that is fair to be honest.

  2. Marchelewska Karolina says:

    I personally think people should not be hired regard only to what they post online. Sometimes people get their social media profiles hacked and do not have an influence what the other person is doing with their account. In case of such situation, how it should be considered by program? As this leaves permanent traces in social media profiles, the program would consider all of the online actions as the person which is assessed and this could be in many cases harmful for the potential worker.

Leave a Reply

WordPress › Error

There has been a critical error on this website.

Learn more about debugging in WordPress.