The Complexity of Bringing Artificial Intelligence to the Workplace
It’s a warm, late spring day and you leap out of your office chair headed for lunch with your co-workers. It’s perfect for a patio lunch. Best of all, you can take a little extra time because you got the major report and dashboard done in just under three hours. That work used to take you two full workdays! Benson, your new personal Artificial Intelligence (AI) bot cut that time down to a few hours.
That may well be where we eventually get when it comes to AI in the workplace. It’s also quite cool, for some scary and for others liberating. To get to that point though, is going to be a bigger challenge than you might think.
Artificial Intelligence is an umbrella term that includes Natural Language Processing, Neural Networks, Machine Learning and some other tools. All applications of AI today are what are known as “narrow.” In that they’re very good, but at one specific job. Some AI tools are getting better at doing more than one task, but they’re a very long way off from becoming personalised for each employee or being more generalized for multi-tasking in the corporate world.
There are several challenges to be overcome before we get to the state described in the first paragraph. AI needs a lot of training. It also needs a lot of data and requires specialized chips for processing at scale and consumes a vast amount of energy. You may think, if you use Alexa or Siri, that it’s all just done on your smartphone or smartspeaker. Apple, Amazon and Google have moved a lot of the processing onto their devices using these specialized chips, but not all of it. Some of those seemingly simple requests have to go through a massive data centre somewhere, where all that data that an AI engine needs is stored and accessible at scale.
And that brings us to the data required for an AI bot or engine to be useful to a corporation, law firm or healthcare setting. This means that a company will have to employ excellent data governance and management. Something 97% of business struggle with today. Think about your own organisation and how much do you struggle to get the right data and how much do you trust the data you get? With many larger organisations, getting access to data can be an issue in and of itself, sometimes taking weeks or even months to get the data you need.
One way this is being dealt with is companies moving their data to the cloud on platforms like Microsoft’s Azure or Amazon’s AWS. Some move their data to the “edge”, which is called edge computing, where it’s closer to the cloud for faster and more economical processing. Even if these services are used it still means good data management is required. Your own data management needs to be excellent. This may well be the biggest challenge for organisations to leverage and create highly customized AI tools for their organisation. Many businesses still see data storage, management and governance as a cost centre. They should treat their data as an asset. That’s called infonomics.
The other challenge is the governance of using Artificial Intelligence in the workplace. Issues such as human rights, protecting human agency, racial and gender bias, the rules around using personal data from say customers or patients. Plus the ethics relating to internal employees, contractors and market competition.
There is a fairly large number of AI tools in businesses today and they’re proving very effective. From customer service applications to mundane processing uses. Small businesses are now able to take advantage of AI tools as well, through using Azure or AWS services where they can access pooled data and leverage their own data.
Eventually, we may get there, but there are a lot of challenges to work through. AI may start to advance at an exponential rate, likely it will. But as with most technologies, it’s more about the humans. How we decide to use the technology and the ability of organisations to understand how to use it effectively. Some management may resist it because they don’t fully understand it, fear the cultural impacts and aren’t able to comprehend the impact of the changes on the organisation. So. It’s complicated.
