ChatGPT in the workplace by citrus HR

ChatGPT in the workplace by citrus HR

We all know how quickly technology is advancing and with each great leap in artificial intelligence (AI), comes more questions. AI is part of our day-to-day lives now, which means we’re still learning how it impacts our working world and the jobs in it. ChatGPT is the latest hot topic in AI innovation and we’re here to give you the basics, so you can understand a little more about this new technology and how it might be used in your workplace.

What is it?

ChatGPT is what’s known as a chatbot. Chatbots are computer programs that are designed to respond like a human when you interact with them. You might have seen or talked to them on websites. The idea is that you can have a conversation with ChatGPT: it can answer questions, admit its own mistake and even perform writing tasks for you.

How does it work?

Its knowledge is taken from data gathered from the internet. The bot will use this information to answer questions you ask it and has been programmed to understand what sort of response a human would expect.

How might it be used in the workplace?

Chatbots can be quite fun to play around with, which is a great way of finding out what they can do for you. We’ve also made a list of suggestions for how this AI could be used to your advantage.

Research

ChatGPT can research topics very quickly and is great at presenting the results in an easy to read format.

Drafting

One of its main skills is writing – the chatbot can draft material that sounds like it was created by a human. ChatGPT could be used to writing marketing materials, essays, and even computer code.

Recruitment

AI, in broader terms, can process a job applicant’s data (which might include CVs, questionnaires, and interviews) to help you find suitable applicant.

Customer service

Newer AI can respond to customers’ queries in a much more human-like way than traditional AI tools. It’s much less ‘robot’.

Energy supplier Octopus recently reported that it had introduced AI to respond to customers’ email queries. The AI is doing the work of 250 people with a higher user satisfaction rate (80%) than human call operators (65%) – although it’s worth noting that humans are used to sense-check replies.

What are the issues to consider?

Of course, like anything else, ChatGPT has limitations and things to be cautious of (especially when using it in the workplace.)

Accuracy

It can only answer based on the data is has been trained on. As this data was gathered at the end 2021, there are already gaps in ChatGPT’s knowledge. The makers themselves admit “ChatGPT sometimes writes plausible sounding but incorrect or nonsensical answers”. So be careful if using it for recent research!

Bias

AI only knows what it knows. What this means is that if an AI is biased, it will continue to reinforce these biases. Not great for equality and diversity!

For example, in 2018 Amazon had to withdraw some AI it used in recruitment in the USA because it appeared to enforce the company’s existing ten-year practice of selecting predominantly men.

Breach of Intellectual Property (IP) rights

It’s clear that ChatGPT was trained on a large amount of data, much of which was from the internet. At least some of this material will be subject to copyright, but there is no easy way of knowing what material protected by copyright might be included in any material that is generated.

The makers themselves give no warranty that material is not subject to others’ IP rights, so businesses need to be very careful and check outputs do not breach IP rights before publishing anything the bot creates.

Confidentiality and data protection issues

When users enter data, ChatGPT by default adds this data to its database and has the right to use it again for other users – the terms and conditions make this clear, although it is possible to “opt out”. If the data you enter is personal data, then passing it on in this way is likely to be a breach of GDPR – unless there was full transparency when the data was collected.

Because of this, it’s important to be aware that responses generated by ChatGPT may also contain personal data that has not been processed in accordance with GDPR.

So, what should we do about it?

It’s clear that chatbots and AI have both pros and cons in a working environment. It’s also clear that many issues you could face when using AI are already subject to regulation and/or workplace policies and contractual obligations, such as: Confidentiality, Intellectual Property and Data Protection.

In short, what this means is if you’re thinking about using chatbots or AI, you should always be:

• Taking full responsibility when using anything created by or decisions made by AI

• Understanding of the potential drawbacks this technology has, and be cautious of that

• Checking results for accuracy, originality, breach of IP rights and possible personal data implications (don’t take any risks!)

• Putting safeguards in place to ensure any personal data that you control is not misused

• Taking care that anything you publish it is not in breach of GDPR.

The ChatGPT trial has not only showcased the capability of new technology, but also its potential pitfalls and shortcomings. Even the developers themselves are urging governments to consider its impact and how the technology should be regulated. So, while ChatGPT can be a quick way to take care of certain tasks, it’s clear that any use of it in the workplace should be very carefully considered. The more informed you are on the subject, the better a decision you can make.

If you do decide that ChatGPT is the way forward for you in your workplace, we recommend having a policy on how to make sure your team are using it appropriately.