So, here’s the thing. I’m a Trekkie. You know, one of those old school Star Trek nerds who cut her teeth on Kirk, Spock, Scottie and Bones. And yes, I can still remember how many quatloos the Gamesters of Triskelion were willing to pay for the group.
So, I would say I’m pretty used to having quote The Computer unqote, always omnipresent and managing the background tasks. After all – you don’t think Scottie was the one changing the oil on the Enterprise – right?
But, I have to admit – the recent onslaught of Artificial Intelligence applications, and the glut of news stories around our descent into the Matrix – well, it has certainly given me reason for pause. Don’t get me wrong – having the Enterprise computer around sounds like a really handy thing. But I just didn’t really think about the road from here to there – ‘there’ being all the kinks worked out, and confidence that the computer was always going to be on our side, and absolute trust in it’s ability to provide fair, unbiased information.
Well folks – we’re not there…not yet anyway. Today’s AI can be incredibly helpful. And I do agree that while it may take the place of some jobs, it will open opportunities for other jobs we haven’t even considered. AI Prompt Developer anyone?
But again, in it’s current state of evolution (that of a offputtingly powerful, toddler), we’re going to have to be incredibly careful. Sure there are societal questions – but to be even more specific, we as leaders in our companies are going to have to get in front of the wave, and figure out how, when and where we can let AI in.
So, before you (or your team members) start using ChatGPT for everything under the sun – consider these top 4 issues with AI in the workplace.
OK, so the biggest one is confidentiality. I bet you didn’t know that all your ‘chats’ with ChatGPT are stored. No, of course you didn’t know – because most of us just blow past the opening statements and notices on websites – I know I do. Your queries and responses are used to help further teach the tool, but while that’s happening, your stuff has now become available to the public. Just let that soak in for a few minutes.
So, you think to yourself – I need a marketing strategy for my new product, purple glitter nose covers. I’ll just enter the specifics of the product, my target market, and any other relevant information into ChatGPT, and ask it for a marketing strategy. And voila – magically, in about 2 minutes, I have a fully fleshed out marketing strategy for purple glitter nose covers, complete with social media posts, eMail copy – the works. But – you have also just outted your new product to the world. Or, maybe it wasn’t even you – maybe it was your junior marketing coordinator.
According to CyberHaven, a leading data security company: “The average company leaks confidential material to ChatGPT hundreds of times weekly. ChatGPT is incorporating that material into its publicly available knowledge base and sharing it.”
Now, you can opt out of it using your information for learning, and possible pubic exposure. But you have to choose to opt out; and again…..we usually miss those types of notices.
And don’t forget the possible issue of destroying trade secret status. I think Duke Law School said it best: “A trade secret can be protected for an unlimited period of time, unless it is discovered or legally acquired by others and disclosed to the public.” So, that sounds like fun – right. You, as the business owner, thinks something is a trade secret – but it turns out it was put into an AI 3 months ago, and has been available to the public all that time. So – poof – no trade secret protection.
Now, I will (again) mention here that I’m not an attorney – and you should always check with your attorney. But, this all seems like it could potentially be a huge problem.
Alright – I’ll stop with the confidentiality issues. Let’s move on to possible security issues.
While it’s true that the biggest player on the field – ChatGPT – doesn’t have access to the internet, and so wouldn’t interact with web based applications. But the part of ChatGPT that is incorporated into Bing search does. And on top of that – ChatGPT has announced integrations. So developers are falling all over themselves, creating applications that work with ChatGPT and can provide more access to other applications – applications you use to run your business.
For instance, there’s CodyAI. It’s called the AI Trained on Your Business. You train this AI assistant on your business, your team, your processes and your clients with your own knowledgebase. That means uploading information on your business, into the AI. Now, don’t get me wrong – it sounds like the use case could be wildly valuable, but let’s just say – you probably need to carefully vet what’s going into the platform.
“… In addition to being an amazing tool; capable of massively increasing your productivity, providing large amounts of valuable research, and sparking ideas as a brainstorming partner….it can also cripple your company, if you don’t use it thoughtfully.“
And then there’s the problem of AI hallucinations. That’s the industry adopted phrase for – well, sometimes artificial intelligence just lies. Oh, it does it with great authority; and to someone untrained in the subject at hand, it may sound absolutely correct. And still be a flat out lie.
Here’s a recent exchange I had with ChatGPT:
Chat GPT: non-exempt employees in California must also accrue at least two weeks of vacation time per year..
Me: Why do you say non-exempt employees in California must .. accrue at least two weeks of vacation time per year
ChatGPT: I apologize for the error in my previous response. In California, there is no state law requiring employers to provide vacation time to their employees…
What can I say…I guess it was a hallucination.
Now to be fair, just like the confidentiality warning, ChatGPT says, right up front, that it may provide inaccurate information. But again, we tend to blow past those types of warnings – because we’re in a hurry to play with the ‘thing’.
And finally there’s the issue of machine bias. There’s a big scientific explanation for machine bias, but I’ll boil it down into the explanation provided in the White House’s blueprint for an AI Bill of Rights whitepaper – it says, you should not face discrimination by algorithms, and systems should be used and designed in an equitable way.
Don’t waste another minute trying to figure out how to get your business going and what should come next.
Grab your Free 3-Step Leader’s Journey Business Builder course, and get freedom and peace of mind, knowing you are on the right, proven track.
Now I won’t get into the myriad of ways this can play out – but just remember that these AI tools are programmed by people – and people are famous for having biased perspectives and limited exposure. Additionally, AI learns based on the knowledge people have provided. After all, ChatGPT learned basically by scraping information from the internet. And of course, we all know that everything on the internet is true…right?
So, I’ll land the plane by saying – AI is beginning to get lots of attention, from the White House, to Homeland Security all the way down to the states. In fact New York is the first state to create a set of rules around using Automated Employment Decision Tools…yep, a fancy name for Recruiting AI.
So, as an employer – you should really think about the following:
And yes, I use, and actually enjoy using, AI. But only with very specific limits in place – because, well – we aren’t on the Enterprise – quite yet…
Oh, and by the way – I did think about having ChatGPT help me in writing this episode – but I thought better of it.
I don’t want him to get angry.
I wouldn’t like him when he’s angry….