Concerns about the risks AI will bring to their business is cited as the primary barrier most HR leaders have in introducing AI into the people function.
We’ve successfully deployed our HR AI Assistant for global companies like Smith & Nephew and have seen HR service requests reduce by up to 75% as a result!
We help our clients mitigate the common risks raised by HR, Legal and IT Teams, here’s how.
#1 worry for HR leaders: ‘Your new AI assistant may access sensitive personal data, increasing the risk of data breaches and non-compliance with privacy regulations.’
Mitigation:
Take baby steps and choose a first project where personal data is kept to a minimum. Our AI Assistant can give vastly improved answers with simple non-sensitive data like job title, country and department.
Work with your vendor to understand where they store your data, which 3rd parties they pass it to, how the encryption works and more. If this sounds technical, don’t worry, this is where your vendor will work with your information security and IT teams to ensure your data is safe. To get ahead of the curve, ask your vendor for an Infosec Pack up front; any credible vendor will have one ready to send. Here’s Applaud’s.
‘AI assistants may deliver inaccurate information due to outdated or missing policy information or being unable to personalize the responses based on your employment data.’
Mitigation:
Choose an AI assistant that allows your employees to ‘rate’ answers, such as indicating when an answer is incorrect or unhelpful. Continuously monitor the feedback and implement a ‘Business as Usual’ process where ‘bad’ answers are addressed by refreshing, correcting or creating new content. (Our solutions can help with this too!)
An additional problem we often see is with global companies that have local policies. Ensure your chosen AI assistant recognises where your employees work and can use this information to deliver the relevant and correct policies and content for the location or department they are working in. This greatly reduces vague or inaccurate information.
‘Employees and management may resist adopting AI assistants due to fear of job loss, lack of trust in the technology, or fatigue with unhelpful ‘old style’ chatbots.’
Mitigation:
Involve key stakeholder groups in your project; widen your testing out to regular employees and managers in the business who would not normally ‘test’ HR IT projects.
Build a set of internal champions.
Any AI project should involve a well-crafted change management program to educate employees about your new tool. Build excitement and curiosity around how this initiative will elevate your HR Service Delivery and deliver answers to employees faster. Create internal email blasts, posters, intranet alerts and email signatures to promote your new tool.
Highlight AI’s role as a tool for augmentation rather than replacement and focus on the new opportunities this will bring to your HR team, for example, there will be new tasks to monitor the quality of conversations taking place and a greater focus on knowledge management.
‘AI systems can be seen as "black boxes," making it difficult to understand or explain their answers, which can lead to mistrust and compliance issues. It can also make problem-solving difficult when incorrect answers are provided.’
Mitigation:
When choosing a solution from a vendor, make sure there is a facility to ‘debug’ any answers that come back. For example, Applaud’s AI Assistant can be asked to provide the information / document sources and references it used for any answer, giving our customers a mechanism to trace a ‘bad answer’ back to the documentation that the AI used to provide that answer. From here you can correct the source content (the problem) rather than the responses in the chatbot (the symptom!).
Get your vendor to help you provide clear documentation on how AI decisions are made and communicate this to your stakeholders as part of your change management process.
‘HR teams may lack the skills required to manage, interpret, and leverage AI tools effectively, leading to underutilization of the technology once the initial deployment project has been completed.’
Mitigation:
Invest in training and up-skill HR personnel in AI literacy. There is a wealth of free training available to explain AI concepts like Prompt Engineering, Hallucinations and Language Models. Get your HR team to create accounts with free tools like ChatGPT, where there are free resources available and where the AI can answer questions about itself.
Your vendor should also provide a good handover at the end of the deployment project, enabling your HRIS team to be able to extend the solution (for example, widening out the pool of topics AI can answer questions on). Be sure to audit the responses, report on usage and problem-solve non-optimal responses.
"The future of HR is AI-powered and any HR practitioner who does not become educated and familiar with AI concepts will be left behind."
Ivan Harding, CEO and Co-Founder of Applaud
Deploying a new HR AI assistant is a terrific way to deliver immediate value to your employees but also for your HR team to up-skill.
However you choose to dip your toes into the world of AI, there is one thing we know for sure and that's the future of HR is AI-powered and any HR practitioner who does not become educated and familiar with AI concepts will be left behind.