Embracing AI in the workplace

July 10, 2023

By Thomas McLachlan, Head of Aprio Digital

The AI (r)evolution looks like it’s here to stay. Unless governments are able to hit the pause button soon, we’d better get our heads around the fact that there’s a new digital employee in the office.

What’s it all about?

We’ve all heard it all already, but here’s just a quick recap – Chat GPT is an example of something called Generative AI. In fact, it’s actually subset category called a large language model (LLM), which generates novel combinations of text in the form of natural-sounding language. It really stands out from traditional chatbots because it can provide a truly human-like level of interaction and expertise across a wide range of topics. As a result, LLMs have the potential to revolutionise how we approach daily tasks, improving efficiency, utility, and engagement. So it’s no surprise that, as this JP Morgan economist points out, Big Tech has announced multi-year investment commitments to get in early on Generative AI.

You can’t stop a good thing (but can we at least slow it down?)

The good news is that we’re all stumbling across this tech pretty much at the same time, the bad news is it’s moving in fast. A city in Japan is already using Chat GPT for a number of its administrative tasks, while Morgan Stanley used GPT-4 (Chat GPT’s latest version at the time of publishing this piece) to create a chatbot that can access, assess and analyse its entire knowledge library and help deliver quick and easy info and content to financial advisors. The accolades keep on coming – in February it had already passed a Google coding job interview for a Level 3 engineer, with a $183,000 salary.

In November of 2022, OpenAI released the GPT-3.5, a public prototype that quickly gained popularity. Within five days, it had attracted one million users, and by January 2023, it had surpassed 100 million users, becoming the fastest-growing platform ever.

Preparing for the future workforce

While not everyone knows how to use Chat GPT, at some point, they will. Employees may want to try it out to see what it can do, and that’s where things could get tricky. Here are some potential issues:

• Ethical concerns: There are some very valid concerns about the ethical implications of AI, such as how it will impact job displacement, the level of bias in its algorithms, and the massive data privacy issues it carries.

• Workforce reskilling: AI adoption may necessitate retraining and reskilling of employees to adapt to new roles and technologies.

• Resistance to change: Employees may feel threatened by AI adoption, fearing job loss or diminished importance in the workplace.

• Infrastructure and investment: Implementing AI in the workplace requires significant investment in infrastructure, data management, and employee training.

Mitigating risk

As AI continues to transform the workplace, it is crucial for businesses to prepare for the future workforce by taking the following steps.

Invest in employee training and development

Investing in employee training and development will become essential for businesses to stay competitive in an AI-driven world but it’s also the right thing to do. As a new technology it’s likely to cause discomfort, so let’s open the lid and see what’s inside. Continuous learning opportunities should be provided to help employees adapt to new roles, technologies, and skill sets required in the age of AI.

Encourage collaboration and open communication

Instead of blacklisting or threatening employees for using AI, create a safe space for experimentation. Education and collaboration are key to ensuring employees use the technology safely. Promote knowledge-sharing and innovation sessions and consider getting an expert to train employees on risks and opportunities. AI can even enhance discussion and collaboration in the workplace.

Emphasise human skills – especially creativity

While technical skills are crucial, the importance of human skills such as creativity, critical thinking, empathy, and emotional intelligence cannot be overstated. There’s already a thriving class of semi-technical experts calling themselves “prompt engineers’. Their core speciality is designing ways to – essentially – design reusable prompts to help you get the best out of Chat GPT. In other words, they’re early AI whisperers.

Businesses may need to start actively cultivating these skills within their workforce, and encouraging employees to get creative in how they use the tech – not everyone believes they are creative but they truly are. You wouldn’t have survived this long if that wasn’t the case. And this creativity is what will (hopefully) remain in demand even as AI takes on more tasks.

Don’t be reckless with sensitive information

Clients trust you with their information. It’s vital that your employees know this and understand what they can and can’t share on a platform like Chat GPT. Make sure you don’t violate any non-disclosure agreements (NDA) or privacy policies, and consider these steps, at a minimum:

• First, have a conversation with your employees – they are already messing around with the technology. Allow them to be open about it and share how they’re using it. It’s the best way to know whether there may be risks.

• Also review your NDA and company policies to understand how to address confidentiality with third parties (like Chat GPT). As a third party, you may be restricted from sharing certain content, legally.

• Chat GPT uses your search prompts to help train its AI. If you don’t want your data used for training (and if it’s sensitive you probably don’t), then you need to make sure to opt-out on the Chat GPT platform.

• Consult with your company’s legal or privacy team, if available, before using Chat GPT.

• Ensure that any confidential information is removed from any documents prior to using them with Chat GPT by either deleting it or replacing it with false information.Top of Form

• Consider updating your policies and contracts to realistically reflect how you will be using the technology, as soon as possible. Bear in mind that it’s likely to change a good few times overthe next few years.

To conclude, I feel it’s only appropriate to answer the question “Did we use AI to write this article?”

Of course we did. Well, to be more accurate –we came up with the content and topic guidelines, then we asked AI for an outline. We then filled in the gaps (because we love writing) and when it was all done, we asked Chat GPT to give the article a once over and make sure the tone was consistent. To be honest, we only actually implemented two of its updates which is why the article is so long. Can you spot what they are?

We’ve even trained it to write like Thomas, using some of his previous content. So, there you have it.