The Rising use of CHATGPT and its implications

So, what is CHATGPT and what does it do?

CHATGPT is described as interacting with you in a conversational way. It is based upon a large dataset of text with the aim of developing human like responses. Open AI, its developers acknowledge that CHATGPT has a number of limitations, including:

(i)               providing plausible sounding but incorrect nonsensical answers;

(ii)              being sensitive to tweaks to the way a question is asked

(iii)            being excessively verbose

 

Having acknowledged that its bot has limitations what risks does CHATGPT pose for businesses and HR Teams.

Knowing whether CHATGPT or any other similar bot is being used to create content is important. Without knowledge that such bots are being used employers are in a difficult position to know what risks are being created by their employees.

One of those risks is the use of other’s data or copyrighted material. CHATGPT will produce and create documents which will be produced from information that is stored and held on the internet.  Some of that information may include material which is subject to copyright.  The difficulty for the business is that it may not know that it has, or is even likely to have, infringed another’s copyright, if it does not know that CHATGPT is being used. Even if it is aware of its use, it is unlikely to know whether any material used by CHATGPT is subject to copyright, because it will not know what source information has been used.  

Equally, CHATGPT relies upon data. The Open AI terms of use provide that users agree that CHATGPT can use any input data and the product produced to “develop and improve” the system unless a specific opt out is utilised. The immediate risk that arises is that the opt-out is not put into effect, but more importantly that confidential information is input into CHATGPT which then loses its confidentiality because it is available to the bot and others who use it. The input data becomes part of the bot’s stored knowledge, the fact that confidential information may have been “lost” in this manner, will not be known immediately or possibly for some time. Where it is the employer’s own confidential data that is one matter, as it may damage the employer’s business directly or because it makes it more difficult, if not impossible, to prevent employee competition in breach of an ex-employee’s contract. If it is a third party’s confidential data, then the employer may become liable to potentially significant claims for damages.

As noted above, Open AI warn on their website that the information produced may not be accurate as it can provide “plausible sounding but incorrect or nonsensical answers”. Humans can, of course do that too, but if the bot is being used as a short-cut, those using it may well decide that the answer produced is accurate, without any further investigation or enquiry. More importantly, CHATGPT may not be fully up-to-date as it can only work on and use information that it has access to, which may mean that some of its responses, lack the most recent information, necessitating further work by those posing questions to it.

Another issue arising out of the use of AI is the development of junior employees. AI of all types makes life quicker and more efficient. Undertaking research and constructing arguments and theories is part of a developmental process that all employees, until now, have needed to take. If those opportunities are reduced or removed, will that impede the development of the future business leaders. A careful choice will need to be made if CHATGPT is to be used as to the circumstances in which it should be used.

HR Departments need to be aware of the use of CHATGPT not only because it can be used in the workplace, but it can also be used to assist with the drafting of application letters and CVs, potentially permitting candidates to create a misleading impression of their skills and talents. The need for a personal interaction may therefore be even greater. HR teams have already had to wrestle with the difficulties of bias arising from the way in which AI programmes have been written or produced. CHATGPT as another AI platform will not be immune to that problem either, and the output may depend upon how a particular question is phrased.

 

What Should Employers do?

CHATGPT and associated AI Bots are here to stay. Some schools are moving the writing of homework essays to the classroom to avoid the issues caused by CHATGPT. Employers need to consider the use of CHATGPT, many are doing so already but only permitting its use internally, preventing it being used for the giving of advice or in any public facing forum, whilst they assess its use and the risks posed by the use of AI.

Pretending that CHATGPT is not being used in the workplace is pointless; it has been downloaded millions of times, it is being talked about and people are curious about it. Businesses need to consider the risks and benefits of using this particular type of AI and ensure that they have clearly communicated their position on its use to employees.     

 

Call to action

If you require legal advice, our Employment team can assist today. We understand the pressures HR and management teams face and we offer a complete solution to your day-to-day or larger strategic HR issues. Please get in touch on [email protected].