Artificial intelligence (AI) and machine learning have been making their way into pretty much every single industry, as the ability for companies to track their progress and analyze data at the push of a button certainly has mass appeal for both its convenience and potential to save money. In hospitals, robot assisted surgeries are becoming more and more sought after; in construction, predictive analytics are preventing injuries; and even in sports, coaches are using simulations based on artificial intelligence to determine good decisions at given points during a game or match.
The general view on AI is that implementing it where appropriate (time saving, money saving) is almost always a good idea regarding the bottom line, but the implementation itself tends to cause some issues within companies, especially larger ones. Here are # challenges that companies face when implementing AI measures.
According to Andrew Schwarz, PhD, professor in the dept. of entrepreneurship and information systems at Louisiana State University Online, the future of business is clearly being disrupted by the leveraging of Artificial Intelligence. But this will not be without challenges. Companies are seeking to meet the expectations of their stakeholders while balancing the ethical challenges that AI will confront them with, including the potential impacts on workforce, decision making, accountability structures, and biases in the underlying data used to train the machine learning algorithms.
Beyond stakeholders and ethics, these new technologies will need to integrate securely into already existing processes and legacy systems, while delivering an output that is trustworthy and transparent. To overcome these challenges, companies need to start slow, innovate iteratively, and put an AI strategy in place with the right people, the right process, and the right technology that will enable the digital transformation that AI can create, adds Dr. Schwarz.
Bias
Can a machine make a decision based on favoritism for one group of people over another? If it’s programmed that way, it can, and because of this, many people balk at AI because fears about biases that may come with it. In human resources, for instance, there are many AI programs available to make sorting through digital resumes much, much easier. There have actually been cases where data would suggest that a machine (or more directly, the people who programmed it) made it easier for a certain demographic to get hired compared to others.
In addition to programming issues, machines can also learn bias if the information they are being given to analyze only comes from a certain group of individuals. A good example of this would be a machine analyzing a survey in Men’s Health and then thinking that all of the people taking the survey were from different genders and races, when it reality, it was probably a very large majority of males taking the survey meaning determinations like 2 out of 3 people have short hair when it reality, it’s probably much less than that, but the survey was taken by primarily men.
A Lack of Understanding
It takes a team effort to implement AI advances in the workplace, but before effort can be put in, understanding has to happen. With this, if you don’t have a robust training program before your implementation, you are setting your team up for failure. There are several ways to do this successfully, with the easiest (yet more expensive) being a third-party training service. In-house programs are also generally effective when the team is onboard with the implementation.
Unfortunately, due to stigmas surrounding robots and other somewhat silly thoughts against AI, some team members might simply never be onboard, and this will be an issue throughout the implementation process, and beyond.
Legal Issues
Legalities surrounding any newer technology are going to be a bit murky, or, at least, somewhat open for interpretation, given the novelty of a given tech advancement. Laws are changing frequently, so teams need to be prepared to implement more changes to adhere to those laws, and they are often related to safety in the workplace that could be jeopardized because of AI.
Ensuring your team has SOPs in place that adhere to laws, and checking up on those laws and changing those SOPs accordingly will keep you out of legal trouble.
Be Confident
Ultimately, even if you’re a bit nervous that some of your team members are under-prepared for AI implementation, it’s a move that needs to be made in many cases. Adaptation after implementation, even for those team members the most against AI, is pretty much unavoidable, so everyone will be on board sooner than later.
