Using AI in the Workplace Comes With Risks
Generative artificial intelligence (AI) tools such as ChatGPT and Claude can help employees draft content, summarise information, brainstorm ideas and complete everyday tasks. But using AI in the workplace also creates risks for employers.
An employee may unintentionally disclose confidential information, rely on inaccurate AI-generated advice or use AI in a way that creates privacy, discrimination or workplace relations concerns.
AI can be a useful workplace tool, but it does not remove the employer's responsibility to protect information, make fair decisions or comply with workplace laws.
Privacy and confidential information
One of the most immediate workplace risks is what employees enter into an AI tool.
An employee might paste an email, upload a document or enter information about a client, employee or workplace matter without considering where that information will go or how it may be handled by the AI service.
This could include:
employee or client personal information
health or other sensitive information
confidential employment or investigation information
commercially sensitive information
passwords or other security information.
Privacy obligations will depend on the organisation, the information involved and how the AI tool is configured and used. Employers should not assume that an AI tool is appropriate for confidential information simply because it is widely used.
Workplace privacy policies, confidentiality requirements and information-handling practices should reflect the use of AI where relevant.
Read Privacy in the Workplace: What Employers Need to Know >
AI can produce inaccurate information
AI-generated content can sound confident and authoritative while still being wrong.
This is a particularly important consideration when AI is used for workplace relations, employment contracts, policies or other HR matters.
An AI tool may:
provide outdated information
misunderstand Australian workplace laws
invent or misrepresent sources
produce an incomplete answer
apply information from another jurisdiction.
AI-generated workplace advice should be treated as information that requires appropriate human checking, not as a substitute for reliable workplace relations advice.
For employers, relying on an incorrect AI-generated answer could result in inappropriate workplace processes, inaccurate employment documentation or compliance risks.
The responsibility for a decision does not transfer to the AI tool simply because the tool produced the information.
Bias and discrimination
AI systems can also create risks around bias and discrimination.
AI outputs may reflect biases present in the data, systems or instructions used to generate them. The risk becomes particularly significant where AI is used to assist with decisions about people.
For example, employers should carefully consider the use of AI in:
recruitment and candidate screening
performance assessment
promotion or development decisions
employee monitoring
disciplinary or termination decisions.
An employer should not assume that an AI-generated recommendation is objective simply because it was produced by technology.
Human oversight remains crucial when AI is used in processes that affect employees or candidates.
AI and workplace relations compliance
AI can generate employment contracts, policies, letters, performance documents and other HR material in seconds. That does not mean the material is suitable for use.
Workplace laws and industrial instruments can be complex and change over time. An AI-generated document may contain incorrect legal references, omit important requirements or fail to reflect the circumstances of a particular workplace.
Employers should ensure that important workplace decisions and documents receive appropriate human review and, where necessary, seek professional advice.
Employees need clear expectations about AI
Simply telling employees to “use AI responsibly” may not be enough.
Employees need to understand what their employer expects when using AI at work, including expectations around:
confidential and personal information
approved and unapproved AI tools
checking AI-generated information
use of AI in work products
human review of AI-generated content
workplace conduct and discrimination
intellectual property and confidential business information.
The appropriate boundaries will depend on how an organisation uses AI and the types of information employees handle.
A policy that was written before generative AI became part of everyday workplace activity may no longer address the risks employees are encountering.
AI is evolving. Workplace practices need to keep up.
AI tools are changing quickly, and employers may not be able to predict every way employees will use them.
The practical response is not necessarily to prohibit every AI tool. Instead, employers should understand where AI is being used, identify the workplace risks and establish clear expectations that fit their organisation.
Good AI governance starts with knowing what your people are using AI for and putting appropriate boundaries around its use.
This can form part of broader workplace policies covering privacy, confidentiality, acceptable technology use, information security and employee conduct.
How Workplace Plus can help
AI can create genuine efficiencies for employers, but it also introduces new workplace risks that need to be considered alongside existing HR and compliance obligations.
Workplace Plus can help employers review workplace policies, documentation and HR practices to ensure they reflect the way work is actually being performed, including the use of new technologies.
For more information, contact us today.
