Fair Credit Reporting Act

For good or bad, the use of AI is already prevalent and its potential uses are expanding quickly, including to the workplace.  LinkedIn is currently suing a competitor, hiQ Labs, for use of information “scraped” from the social network’s site and used for AI analysis.  hiQ uses the information gleaned from LinkedIn to predict whether employee are likely to leave their jobs.  While the issue in the lawsuit is whether outside companies have the right to use information made public on social media sites and does not involve any employment work-place privacy issues, the lawsuit has disclosed how AI is currently being used and in the workplace.  AI is quickly being adopted, and its effects will have huge ramifications for employers and employees.  This Friday’s Five discusses five impacts AI will have in the employment context:

1. Predictions of whether employees are likely to leave their jobs.

The analysis being done by hiQ Labs is a prime example of information that would be highly relevant to employers, employer’s competitors, recruiters and others.  As the Wall Street Journal article notes:

Among its services, hiQ monitors and analyzes LinkedIn profile pages to see who is polishing their résumés and liable to be poached, assigning so-called flight-risk scores to individual employees.

LinkedIn’s primary argument in suing hiQ to stops its “scraping” of LinkedIn’s information is that if LinkedIn users understood that their data was being gathered and used in this manner that they would be reluctant to share information and update their profiles.  This illustrates that there is value in the information being shared on LinkedIn when AI can analyze user’s data.  Regardless of which company has access to it, the fact that LinkedIn is suing over who has access to this data establishes how valuable the data is.  Employers are likely to begin using this data to evaluate their workforce in the near future, if it is not already occurring.

2. Set pay and performance standards.

One positive use of AI in the workplace could be as an aid to highlight good performers in a company and remind the managers to provide positive feedback or raises to high performing employees to increase employee retention.  Another potential use is analyzing data to set pay scales commensurate with the market for a particular locale or skill set.

3. Predictions of potential litigation.

Just as AI has been used to predict future mechanical failures of engines or other devices based on data history and monitoring the device, AI will likely be used to highlight employees who may pose a litigation risk.  Just as hiQ sets flight-risk scores, it is conceivable to set litigation-risk scores based on data.  Not commenting on whether this is appropriate (or legal) to do, the reality is that AI can and will be used for this analysis.

4. Help evaluate candidates interviewing for a job.

AI will likely be used in helping companies evaluate candidates for a job.  AI could evaluate education, experience, and other data obtained through the internet to predict an employee’s likely fit with the company as well as skill set.  There are laws already in place about employer’s use of certain public information, such as credit history and criminal backgrounds that must be excluded from such analysis, employers would have to approach this type of analysis cautiously to ensure compliance with existing laws.

5. Will there be a backlash for use of AI in the employment context, and will it be regulated?

Employers are already regulated on how they can use background information about candidates and employees under the Fair Credit Reporting Act (FCRA) and California’s Investigative Consumer Reporting Agencies Act (ICRAA).  Similarly, AI is using background information known about a person and comparing that data to a wide data set to glean likely future outcomes.  There could be a case made that just as the FCRA and ICRRA create obligations to provide notice to employees about the background information that an employer is relying upon to make an employment decision for the employee to correct any mistakes in that information, employees should be able to see the data being relied upon in the AI analysis.  However, given that AI can gather and process a huge amount of data, it might be impossible to review all of the data.  Moreover, the data relied upon by AI about the employee’s background may be very accurate, but the algorithms relied upon by the AI might weight information in a way that does not result in accurate predictions.  Don’t forget, AI predictions are just that – predictions.  Nevertheless, employers are always looking for a small advantage over competitors, and AI may be one additional tool to do this.  However, like many other areas of technology, the legal system is slow to adopt to technical advancements.  AI in the workplace exists and is being used, employers and the legal system needs to start considering its ethical and legal parameters.

Mat Honan at Gizmodo wrote recently about a new company that helps employers search applicant’s “internet background” to assist in the hiring process. As Mat rightly points out, much of the concern over this “new technology” is overblown, and as he puts it, "[e]mployers would have to be stupid not to Google job candidates."  As I have pointed out before, much of the unduly concern is that lawyers don’t understand the technology, and therefore if they don’t understand it, their client’s use of the technology can only lead to bad things.

I think Guy Kawasaki had a great perspective on this issue when I recently interviewed him. He said he would be worried about a job applicant who did not have a Facebook page: what is wrong with this person? Is he anti-social? Is he not with the times or just simply does not understand simple technology? As Mat points out as well, with some common sense a job applicant can easily manage the results of an online search by being careful about which information he or she provides to the employer. For example, an internet search for the job applicant’s private email address might turn up more personal information than if the applicant has a separate email they only use for work purposes and lists on their c.v.

From the employer’s perspective I don’t think the analysis changes much for searching employees background on the Internet:

Generally, under Federal law, employers may utilize social networking sites to conduct background checks on employees if:

  1. The employer and/or its agents conduct the background check themselves;
  2. The site is readily accessible to the public;
  3. The employer does not need to create a false alias to access the site;
  4. The employer does not have to provide any false information to gain access to the site; and
  5. The employer does not use the information learned from the site in a discriminatory manner or otherwise prohibited by law.