How can Fleetwood Mac help with AI?
Remember when Fleetwood Mac sang, “Don't stop thinking about tomorrow, don't stop, it'll soon be here”? Well, it’s here.
Specifically, artificial intelligence (AI) is here and it's not going anywhere. In fact, it’s only going to get better. While you might approach it from a place of fear, there are a ton of opportunities too.
To start, AI isn’t exactly new. Tech companies have been successfully using software to mimic human behaviour and decision-making for years. Think Siri.
What’s new is the quality of the machine learning – an AI subset. It uses models that analyze massive data sets. Think an entire library in minutes. It finds patterns and logic, then uses that knowledge to refine its algorithms in such a way that a simple command from a user can be understood, interpreted and applied to perform a complex task. Think ChatGPT.
There are a bunch of legal, communications, marketing and other considerations. Let’s delve into just a few before we present a solution for organizations to mitigate those problems and adapt.
Privacy is a big concern. Is a company really going to put its trade secrets, proprietary and confidential information into ChatGPT in order to have it spit out something helpful. Probably not. Conversely, is a company really going to use generative AI for something complex that might involve personal information that the software scraped from the internet? Probably not. Too risky.
Intellectual property infringement is also a problem. The AI software you’re using might be using data that is the intellectual property of others. As a result, the product it spits probably does too. Not ideal.
Copyright presents another issue. Copyright of something that isn’t original is typically not possible. If the AI software produces something for you, it can produce something identical for someone else that inputs the same parameters. It could also be using original works and infringing upon already copyrighted material.
Finally, bias is something that has plagued AI in the past. If the data used for machine learning is biased, the result is going to be biased too. This could have big implications on efforts to advance meaningful diversity, equity and inclusion initiatives.
So where does AI go from here?
ChatGPT has said it’s most recent version is 571-times more powerful than the version it released just months ago. Given the pace of improvement, it’s only going to get better. It’s here to stay so best to adapt and harness its power. But how?
There is plenty of low-hanging fruit. It can easily be used for lower-value, commoditized functions to lower costs and refocus staff on quality assurance and editing. Examples include graphic design, press releases, voice-over work, content creation of all sorts (video, audio, ads, blogs, social media, etc.), sentiment analysis and social listening, ideation and brainstorming, creating business models, policies and plans, and even drafting simple legal documents (though no AI software will accept any responsibility, liability or accountability for that legal help). The list will grow everyday.
Don’t think it’s already happening? Think again. Your company can pay a fee to “hire” a service like Cody and customize the input to essentially create an employee that knows your business. No salary, no pension, no benefits, no statutory deductions, no office furniture. Just a computer and a fee. This will displace many people from their traditional work, especially, in fields like marketing, public relations, communications.
From that challenge also comes a solution. And we turn again to Fleetwood Mac.
Although in a different context, they once sang, “You can go your own way” and it’s something that could apply to AI too.
Instead of relying on these great new AI tools that are fed and trained by the entire world, and therefore present all the problems outlined above, make your own AI system and customize it for your own company.
It solves a variety of problems. The company can control what goes in, thus making sure it can use that data for the intended purpose. It can customize and automate the data input process to improve customer experience and help its AI software can learn constantly and in real-time. This can help alleviate intellectual property infringement concerns, as well as bias issues. It can ensure what the software produces is original and can then be sufficiently edited by humans to ensure copyright can be claimed.
The company can control the dataset, thereby allowing it to train the system to know its key customs, norms and policies, and culture. It can input data about its customers, potential customers, audience segments, competitors, trade secrets, etc. Essentially, instead of garbage in, garbage out, you put gold in and get gold out.
Two caveats. The first is expense. Creating such a system is not initially cheap but the cost will come down over time. For example, think about how much cheaper data storage is now than just a few years ago. Second, the system will have less data so it could theoretically be less smart. It’s not going to know what ChatGPT knows but that’s kind of the point. Do you really want or need it to?
The challenge for those in communications and marketing, among others, will be to avoid lamenting the loss of what once was and embrace what could be. Ignore the instinct to be tactical editors and proofreaders and instead embrace the technology. Be the people who figure out how AI can fit strategically into the business and how the company’s new machine can learn. Learn to set the parameters rather than letting the AI set the parameters of your work. The added benefit for marketing and communications department is that doing so will help show value, decrease lower-value headcount but add higher-value, more strategic headcount.