No other technology today has the level of hype associated with artificial intelligence (AI). Countless news articles, blog posts, videos, etc. have been published predicting everything from drastic changes in the workplace to the complete destruction of humanity. With all this discussion, it can be difficult to determine exactly how, and to what extent, your organization will be affected.
Work product created by AI is fraught with risk. For example, some AI tools can "make up" information (known as "hallucinations") rendering these AI tools unreliable in certain contexts. Further, if AI is trained based on someone else's intellectual property, it may, under some circumstances, produce results that infringe on that intellectual property. Also, information provided to an AI tool, if used to train future versions of the AI, can potentially be extracted, risking the exposure of trade secrets.
With so many unknowns, here are three practical (and reasonable) actions you can take now to ensure your organization is protected.
First, ensure that your organization has a policy in place to govern how AI is to be used or not used. A starting point may be creating a blacklist of AI services that are not permitted. These can include free online AI services where reliability, data privacy policies and AI training methodologies cannot be easily determined. Also, as AI is inevitably incorporated into the software and services that your organization already uses, it may be helpful to implement employee training on how confidential information is to be used, or not used, with these tools and services.
Second, as your own providers add AI tools to their services, review how your current service agreements allocate the additional risks. Ensure that the terms in your service agreements clearly indicate how information provided to the AI may, and may not, be used. Ensure that you have some level of notice and control over updates and changes to the underlying AI that may affect your organization.
Third, develop a rubric for evaluating new AI tools for use in your organization. AI is inherently less predictable than other forms of software. You may wish to evaluate AI tools in the same way you might evaluate a new employee rather than a new productivity tool. What training has the AI received? What are its strengths? Weaknesses? What are its liabilities and work history? Any references?
While no one knows the full effect of AI, taking these steps can position your organization to safely adopt and use the latest in AI technology. At Parsons Behle & Latimer, we have expertise dealing with issues in AI. Let us know how we can help you.
The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.