I partner with online-learning and training site HR Jetpack to deliver courses and webinars on Cybersecurity and Artificial Intelligence topics specifically for HR professionals. In my latest webinar I had a really engaged audience who asked some fantastic questions which are worth sharing.
My organization uses AI to replace in-person interviews and I'm concerned that we might be using AI too much. What are your thoughts on striking that balance?
It is critical to always keep in mind the motivation for why these processes are in place. The point of the interview is to filter and find the best possible candidate. So if the AI is doing that better than a person could, then it is the right tool for the job. If it isn't able to do that yet, then it has been pushed out prematurely.
That happens a lot with technology, either to cut costs or out of excitement and enthusiasm about a revolutionary new technique which might set the organization apart, impress investors, or move the organization in a new direction. Those moves are a mistake if they are ineffective at achieving their core goal. Never lose sight of what they are meant to be accomplishing.
AI applications based on historical data won't predict accurately until enough data is gathered. How do we promote early adoption in those cases?
Go in with both eyes open. Understand that you are essentially agreeing to be a beta tester for something that is yet unproven. You can't be completely reliant upon the results of this new system, so temporarily you will need to run it in parallel with your existing method of accomplishing the same task.
The payoff is that by starting early, getting your team used to working on the new system and integrating it into your workflows, you might be way ahead of the competition by the time you are ready to trust it. There is a time and place to take this kind of gamble; just make sure that it can't backfire on you and that your organization is properly protected against the risks.
How concerned should we be about AI data getting hacked?
We should all be very concerned. Since the dawn of the internet age, attackers have had the advantage. It has always been easier to attack systems than defend them, and that's just been an unfortunate truth of software and the internet so far. Until we see a paradigm shift in the technology landscape, this probably won't change.
Having said that, the potential benefits to an organization in terms of productivity are too great to pass up due to cybersecurity fears. I would expect most organizations to invest in their defensive teams alongside their investment in AI, and there will be a surge of hiring in both fields going forward.
I'm in favor of using AI to save time, but I don't want to lose control over what I do. How far is AI going to encroach into my role?
Control is a fascinating question, and one that is very different at each organization. I have two examples of ways that AI might take "control" as we give those systems more traditionally human decision-making power.
1) Being the Boss: In some offices, the boss is the person who assigns the day's tasks to the team. For the person doing that work, I would ask if they really care who it is that actually assigned them that work? Today, it might be their boss Sally. Tomorrow it might be a software algorithm. The important thing to remember is that job security related to AI is not a totem pole. AI is about tasks — the task which you perform right now, and whether software could replace that task. It's really that simple.
2) Predictive Gambling: For a long time, Wall Street rewarded the best gamblers with big salaries and huge bonuses. Those best able to pick winning stocks were praised for their "gut calls" and "instincts." And now that era is coming to an end. The world's largest investment firms are trusting their own in-house algorithms to invest their billions. Those algorithms make fewer mistakes and have better returns than the people did. Control is being taken away.
At the end of the day, we will always be the ones behind the keyboard, running the operation. However we will increasingly give those systems the ability to run more of the world as they prove they can do it with better results than our own. This is sure to be a controversial dynamic as we move into the future.
