Page 16 - BusinessWest December 9, 2024

“Savvy organizations are also leveraging predictive AI to analyze donor data and gain insights into potential future donors.”
to information, and increasing awareness to drive policy change.
But AI also presents risks that could threaten a non- profit financially, reputationally, and operationally.
How Nonprofits Are Using AI
AI has surged since 2020 thanks to swift advances in technology to generate text, images, and videos. Non- profits are tapping into generative AI and its large lan- guage model (LLM) subset to create text from big sets of data to enhance efficiency and expand their reach. Additionally, nonprofits can use AI to automate repeti- tive tasks, including certain administrative duties like scheduling meetings, data entry, or volunteer manage- ment, so they can instead focus their limited employee and volunteer resources on other important work.
Savvy organizations are also leveraging predictive AI to analyze donor data and gain insights into potential future donors. These insights can guide generative AI to create personalized appeals through targeted commu- nications such as letters, advertising, and other content. Some AI applications are even more ambitious by pro- viding actionable information to people looking to get involved in a cause or mobilize resources.
The Risks in AI — and How to Combat Them
Despite AI’s benefits, risks abound, including errors in word choice, tone, or potential copyright infringe- ment in AI-generated materials. It is critical that orga- nizations have a process to fact-check AI-generated materials and develop usage rules and policies for employees or volunteers supported by awareness train- ing. Organizations should also consider media liability insurance against AI content-related claims of per-
sonal injury, copyright/trademark infringement, and plagiarism.
Cybercrime is another concern. AI has enabled cyber criminals to improve the speed, scale, and auto- mation of cyber attacks. The technology can turbo- charge schemes like phishing or ransomware and be used to mimic voices of real people ‘authorizing’ fraudu- lent activities, known as ‘deepfakes.’
AI systems can be targets as well. If a threat actor was able to compromise a language model and poison the information within it, the outputs generated by AI algorithms leveraging that model could be damaging.
Unfortunately, many nonprofits are resource-chal- lenged and increasingly vulnerable to cyber threats. About 68% of nonprofits have had at least one data breach in the last three years, 75% don’t actively moni- tor their networks, and more than 70% don’t run vulner- ability assessments.
Every organization using or considering AI technol- ogy needs best practices and policies to protect against the potential risks. Here are some steps to consider:
• Document AI use policies. Organizations need to determine who can use public AI tools, and for what purpose. For instance, can business or personal email accounts be linked to the programs? How will access be managed — and by whom?
• Perform due diligence. Third-party AI tools that organizations or its vendors can buy, license, or access cause more than half of all AI failures, which includes providing inaccurate or copyrighted information. Orga- nizations must thoroughly evaluate AI tools and the AI practices of any potential vendors to ensure they are guarding against threats. Rigorous contractual risk
Artificial
Continued on page 19
>>
Let Us Be Your Partners Too.
From the most established partner to the next generation of talent, MBK works with you to understand your objectives, goals, and fiduciary responsibilities.
And we’re here for you all year long— not just during tax season.
413-536-8510 | mbkcpa.com
16 DECEMBER 9, 2024 << INSURANCE >>
BusinessWest