Listen to the podcast.
Traditional Software is Deterministic, but AI is Probabilistic.
Like all software, AI requires developers—but the traditional process of software development doesn’t apply.
“The first step is recognizing that AI is basically software, and so, in some ways it's going to behave like software always has… you're going to have inputs and it's going to go through whatever programming the computer programmer gave it to do, and then it's going to produce an output,” said Robinson. “You have to think about [AI] in that way—that there's still this human element of setting it up and training it and getting it going. That's where the similarities begin to end.”
According to Robinson, instead of being deterministic like traditional software, AI is probabilistic, i.e., the program looks at data and then begins making decisions—a huge step that results in the computer being more human-like.
“That’s where I think the human-machine hybrid that we talked about in the IT Industry Outlook really comes in,” said Robinson. “You're going to be able to get new insights that you might not have seen before from a dataset, but you've got to double check that and make sure that it's not going off in the weeds somewhere.”
Implementing AI Requires IT and Business Units to Collaborate.
As companies continue their digital transformations, IT must collaborate with business units in order to effectively meet goals. But when it comes to AI, over half of companies surveyed by CompTIA felt that AI projects should be handled by the IT team.
“That speaks to the belief that AI is something that happens under the covers, it just needs to get plugged into the architecture or the infrastructure, and then new results are going to come out, and business is going to continue as before but with the new results,” said Robinson. But those results are critical for business units to understand and embrace.
“I think it's really going to be important for the business people to take a look at what the output is, or they know what they want or expect to come out of this in the first place, so be there on the front end and then also be there on the back end of the process,” said April.
AI Comes Down to Data.
Data is central to AI functioning the way it is intended—but that reliance on huge datasets can make AI solutions difficult for companies to implement.
“In the interviews, we saw people saying that one of the biggest hurdles that they're running into is not having the right data to feed into the process, or not being able to collect all of the data that they thought they would be able to,” said Robinson.
Data management has consistently been a hurdle, particularly for small and medium-sized businesses, but when it comes to customizing an experience with AI, data is key.
“You're going to want to feed your own data in to train that product and to get the output that you would want. You can't just rely on some generic training data that has set this product up to be used by you,” said Robinson.
Stay Tuned for More Insights on AI.
CompTIA’s AI research, Emerging Business Opportunities in AI, and the companion paper, Practical Insights on AI, will help businesses more clearly understand the opportunities this emerging technology presents. Listen to the full podcast to find out more about the newest AI research from CompTIA, available soon.