Exceptionally good at administration and highly competent in a wide range of tasks and functions, virtual assistants are ideally suited to taking on a number of tasks without the need for in-depth direction from their clients. However, in the past year or two, we’ve seen another productivity-enhancing development - Artificial Intelligence, otherwise known as AI. Actually, since we’re on the subject of time-saving we'll be using the abbreviations VA (virtual assistant) and AI (artificial intelligence) in this AI guide for VAs.
Although AI is supposedly set to improve various aspects of human existence, there’s no doubt it has the potential to further capabilities in work, and businesses worldwide will be seeking to take advantage of this. The issue, however, is that AI is no replacement for human intelligence - not yet anyway. It still makes mistakes and there are also some potentially serious risks to consider and manage.
Since AI is essentially a processing tool and VAs are responsible for optimising productivity for clients, AI is set to be a game-changing force in the world of virtual assistance. Especially for VAs who are early adopters. However, when incorporating AI into work practices and exploring the various tools and uses, there are a few things you must know and also some potential legal issues to be aware of.
Here are 5 ways virtual assistants can mitigate risk when using AI:
Ensuring Accuracy with AI
AI is not generally recommended as a research tool. An AI platform only has access to information it's been exposed to and, in some circumstances, this information may be misleading, inaccurate, biased or simply out of date. When using a search engine to seek out information, most of us will be screening results and scrutinising sources to ensure reliability. When we instruct AI to conduct research for us we do not always know the sources. Even if these are stated they must be double-checked as AI may misinterpret information or lose context as it repurposes the content, it might even present a work of fiction as a case study because it isn’t always able to spot the difference.
Here are some tips for virtual assistants to ensure accuracy when using AI:
Listen to meeting recordings yourself - AI tools including fireflies.ai and otter.ai can record and transcribe meetings, making it simpler for virtual assistants to prepare notes, and summaries and even create post-meeting actions, as required. However, with the talking over one another, use of humour and sarcasm and jumping between topics that can happen in meetings AI transcripts may not always be completely reliable. Listen to recordings yourself, either before or after checking through the transcript, to ensure everything has been understood and nothing has been overlooked.
Check sources when using AI to create content - Language processing tools such as ChatGPT may use any number of sources to create content and should be able to list them all for you when prompted to. It's important to check all these to ensure you’re satisfied they are a trusted source. Be particularly meticulous in double-checking statistics, case studies and cited research that may harm your professional reputation, or your client's business, if found to be inaccurate.
Test AI Automations - If you set up AI automations, either for yourself, your clients or your client’s customers, then you will need to regularly test these to ensure they’re operating as desired.
Ensuring Quality with AI
Relying solely on AI to produce work - especially ‘original’ content - can be risky. Moreover, it can result in poor quality work for the client because AI tools are incapable of original thought and therefore the content it outputs can be fairly basic and uninspiring without human editing and contribution.
Here are a few ways virtual assistants can still ensure quality when working with AI:
Work WITH AI - The key for virtual assistants is to use AI as your assistant, not your replacement. You are in charge so you give the instructions and you check the work produced to ensure quality control. For example, a grammar and spelling checker, such as Grammarly, will make suggestions for your writing but you are free - and can often be right - to reject these.
Personalise it - The biggest criticism of work produced by AI is that you can tell it’s produced by AI. At least, you often can if it isn’t edited. AI-created content requires human refinement and will usually need to be made to align with the client’s tone of voice and brand guidelines.
Implement fact-checking processes - Before submitting AI-generated work to clients, check for accuracy as emphasised above.
Do your AI Due Diligence
When selecting which AI tools and platforms you may want to harness to improve efficiency, it’s essential to conduct proper research. You’ll want to research the AI’s creator, their track record and reputation, especially concerning security. It can be wise to seek trusted advice and recommendations from both other VAs and also those with some technical knowledge.
Of course, where a client may need you to make use of their AI systems you’ll have less control. When accessing client systems, especially those with artificial intelligence, you should make sure you only ever use it for that particular client and don’t input any personal information. Most importantly, you should obtain a basic knowledge of how each AI tool you use works. Understand its decision-making process and limitations to ensure you can fully harness its potential without overestimating it or trusting it with tasks it is unequipped to properly handle.
Accountability and Liability
When using AI in the course of client work, transparency is key. You should be upfront about when and how you use AI.
To ensure legal compliance, accountability and clarity, VAs using AI should:
Declare the use of AI in terms and conditions and/or client agreements.
Ensure compliance with data protection laws such as the GDPR, Data Protection Act 2018, or other relevant regulations.
Clearly define in contracts who is responsible for errors or issues arising from AI use.
Use secure and encrypted communication channels when handling client data.
Regularly review the terms of service and privacy policies of AI tools to understand how data is processed and stored.
If you are a virtual assistant using AI tools then you might need to update your terms and conditions as well as your privacy policy. To learn more about how to do this, refer to my AI Legal Toolkit which includes a number of AI disclaimers for various uses.
If you’d also like to update your current terms and conditions, I have a legal pack designed solely for VAs you can view here, including terms and conditions, an associate agreement, and all documents you need for your website. Or if you would prefer a solicitor to draft or redraft your terms and conditions please get in touch.
Whilst it may be tempting to downplay or avoid discussing use of AI for your client work, there are a few reasons why it's important to be transparent about AI usage with clients:
Clients may have strict data security policies that rightly prohibit AI usage for certain tasks.
Clients may hold you legally responsible for errors or misleading AI-generated content.
Unauthorised data breaches can lead to legal penalties and reputational damage.
AI providers may store or use input data for training, potentially exposing sensitive information
If your client is based in an EU country then it is a legal requirement to be transparent as per the EU AI Act.
Copyright and Intellectual Property
Copyright for templates, designs and other content you create with AI may not be automatically assigned to you in the same way they are automatically the property of the creator under UK law, because the AI provider may be considered the author, or co-author. This depends not only on the terms and conditions of the AI provider but also on future legislation and regulations. Hence, it will be essential to keep up with any new laws around AI, especially regarding copyright, as the tools themselves and the legislation around them continues to develop. Furthermore, AI-generated content is still in the early stages and there is still a significant risk that the AI may infringe on existing copyrights when generating images, text or designs.
To mitigate this risk, VAs must ensure they check sources, edit AI-generated content and ensure compliance with UK copyright laws. In some cases, you may be able to cite and credit sources. In other circumstances, it may be wiser to create original work rather than risk leaving yourself or your client liable for copyright infringement as a result of using AI.
Best Practice for VAs using AI
Virtual assistants and artificial intelligence should be able to work together to increase productivity and benefit clients, making day-to-day tasks simpler and optimising time and skill sets. While there are AI tools that may help with scheduling, content creation, administration, transcription and data analysis, these tools must be used with knowledge and awareness.
Over-reliance on AI without human oversight can undermine professional credibility and even risk client confidentiality. You must also, as mentioned above, be mindful of data protection and intellectual property.
We are all still learning how to harness the powers of AI but we are also watching as the legal world works out how we will preserve existing rights and freedoms whilst embracing these new technologies. Every person using AI in their professional lives - and particularly VAs who stand to enhance their productivity using AI - must ensure they understand how it works and implement processes for proper usage and compliance.
My AI Legal Toolkit navigates you through helpful checklists and provides necessary template policies and inserts for your legal documents to ensure compliance with data protection and IP laws. Finally, I have a handy 12-point compliance checklist, which you can download for free here.
If you'd like bespoke support on how to use AI legally within your business, please reach out here.
Kommentare