AI in HR: Thoughts on Salesforce's Groundbreaking Approach

What started as a typical interview between "Mad Money" host Jim Cramer and Salesforce CEO Mark Benioff took a bit of a jaw-dropping turn last week when conversations about Salesforce's AI chatbots shifted to Mark stating "We just hired 2 people using AI." He continues to describe this by saying that they used AI to reach out to people, find out if they were happy with their jobs, wanted more money, etc. and when they were interested, it was then passed to a human to continue through the hiring process - simplifying the work for HR recruiters. You can see the full interview below.

My First Reaction

My personal and instinctive reaction was to think about the ethics and harmful potential for such a situation. I imagined AI just reaching out to people on LinkedIn and via direct messaging - regardless of whether they want to be contacted or not. I wondered how they might get the data they needed to determine who to go after. I wondered how those potential recruits would know whether or not they were talking with Salesforce or an imposter. How do we stop bad actors? Just think of all the data bad actors have on us from the many high and low profile breaches there have been! I used to think a breached e-mail address wouldn't be a big deal because all the spam and phishing attempts with the bad grammar and misspellings are pretty standard to stop. But with AI, all of a sudden you have a tool that can speak proper grammar and that spells words correctly! A tool that can mimic and copy images. Yikes!

Allowing For Cognitive Reasoning

After that initial panic settled, I spent a little more time thinking and researching the situation. Not a lot, but just enough to get a little more realistic about what likely really happened. Salesforce likely didn't strike any deals with Microsoft to get access to their LinkedIn data and determine potential candidates. So it's not as likely they were DMing random people. There have also been reports that Salesforce has been focusing on hiring "Boomerang Employees" - former employees who left or were laid off, but interested in coming back to the company. So the likely story is that they used their own data, contacted former employees by the e-mail addresses they still had on record and were working in areas or skillsets they wished to build/rebuild all through automation (AI). It's possible that you could still bring some ethical questions into this, but generally speaking this would be pretty normal if they weren't using AI too. Whew! This isn't as malicious as first imagined and the techie in me feels rather impressed at the innovative thinking to automate efforts that would take some significant human power and time to do!

All's Well That Ends Well?

I'm not here to strike fear into people's hearts. I want to take people through my thought processes with the knowledge and experience I have. There is undoubtedly some really cool things going on here and in the AI world in general. That said, there is absolutely legitimate malicious potential and ethical concerns to be had here. Some, if not all of the scenarios my state of panic imagined are entirely possible - maybe even likely to happen in the future. Data that has been taken from breaches could potentially be used on us. As is normal with technology, bad actors will get better at what they do and it could become harder for us to recognize. Our own knowledge and vigilance to protect against that must continue to grow, unfortunately. Hopefully we will also build tools to prevent and/or help us detect those intrusions. When it comes to AI, ethical use and proper controls will continue to be absolutely paramount. We can't skip this without serious costs and this is why the AI revolution will take many years to fully blossom. It's amazing to hear another business case where automation and intelligence are helping simplify tedious work. I also find it prudent to take those ideas and consider the potential for it - both good and bad - elsewhere in the world.