Question

Describe some of the risks associated with the “bring your own device” movement and some of the policies companies have established to mitigate these risks.

Provide examples of how biases in data sets used to train artificial-intelligence systems resulted in poor performance of these systems across diverse populations.

Answer 1

An increasing number of employers are allowing their employees to use their own personal mobile devices for work purposes a trend known as BYOD (Bring your own device) in fact as many as 80% of smartphones used for work are reportedly employee-owned but this increased usage of personal devices also brings with it some potential pitfalls for employers there are many concerns need to be aware of including:

• Including possible loss of confidential data
• Inappropriate employee postings
• Privacy concerns
• Overtime pay issues
• cell phone use while driving that could lead to employee liability

BYOD is more just a trend it's fast becoming the norm in mobile business communication the move to BYOD brings plenty of advantages for employers some of which might apply to the organization:

• Cost-saving
• Employee comfort
• Portability/connectivity
• Productivity
• Access

employee use of personal devices comes with serious consideration that must be weighed and addressed the most significant of these is of course security loss of private and sensitive data and the possibility of malware wide-scale corporate data breaches make the news almost weekly so this should be utmost priority depending on industry and state in which company may be subject to specific laws governing data breaches protection of sensitive information so the first step before implementing the BYOD policy should be consultation with legal counsel to identify any compliance issues. employees should have secure credentials and passwords to access company servers and information and they should also agree to keep their devices locked and ready to diable in case of loss or theft. there's also a chance that photos or videos on an employees phone could pose a security threat to the company such as images of a product or a design while there are limits on how much your IT department can monitor an employees private device you may restrict the use of these devices for taking pictures for work purposes you should also require employees to use only secure wifi networks and ask them to regularly update their apps and out of the data app on a smartphone can be an open window through which cybercriminal can access the information on the corporate server.

Managing these issues will require that you have solid tech support in place either on staff or by contract IT team can help determine which types of devices will support the activities need and help stay on top of security the sensitive data servers.

finally, if offered the option of BYOD to employees draft a policy that covers what the company expect employees to do it limit the risk to breach as well as the consequences of failure to comply with those requests, for example, refusing the lock of the personal smartphone might mean that the individual would be moved over to a company device the policy would also outline post-employment issues such as how employee access to email data and servers will be shut off what employee leaves the organization and the case the purchase of the devices was subsidized who will retain physical ownership of the device at that time the employee should sign the policy and a copy should in his or her personnel file no policy can limit all risks of a breach via employees personal devices but putting these measures go a long way towards securing private and sensitive data.

employers need a comprehensive BYOD policy that clearly articulates:

• The types of devices the employer permits
• The kind of information an employee may keep on the devices
• What apps are permissible
• What happens if the device is lost or stolen
• All anti-discrimination policies extend to the use of personal devices

Organization need to balance the risks with the benefits despite the many risks of BYOD there are some smart security measures employers can take to minimize liability that there are even some benefits for employers that institute strong policy employes may be responsible with and take better care of their own devices than those provided by an employer, BYOD can also be cost-effective for employers and employees devices are sometimes more sophisticated than those of the employer when employees use their own devices it enables employers to reduce the amount of time needed to train employees to use and provided devices.

AI system can also develop biases and this could be problematic a large company has solicited interviewed and selected candidates over the past ten years some of these candidates join the company this data could be used to train a machine learning system to automatically select the candidates for future postings but if the original data has bias such as hiring disproportionately more men then women or paying men higher salaries for the same jobs then this bias is carried on to the machine learning model future decisions made by the machine learning algorithm will also contain the bias and if those candidates are hired then the bias gets reinforced this is a problem in AI and machine learning that companies need to take a serious look at deep bias data this is not easy because a machine cannot tell the difference between an underlying pattern and a bias in the data.

Algorithms are becoming more and more involved in major decisions in many industries. they are already helping to decide who gets a loan, who is hired or fired, who can travel freely, and even who is arrested and how long they go to jail. if these algorithms are flawed or biased, it could actually amplify injustice and inequality. AI is only as sharp it learns from. for instance, a study at MIT lab found that leading facial recognition systems correctly identified white males' faces 99% of the time, but dark skin female faces, it made mistakes up to 35% of the time. that's likely because the data used to train software to often overwhelmingly white and male, one widely used data set to be more than 75% male and 80% white.