If someone were to ask you about the major themes of 2023, we wouldn’t be surprised if you answered “AI.” With the release of ChatGPT earlier this year, the topic – that has been around for decades – has become top of mind for many people, particularly those of us in tech. Interestingly, when I asked ChatGPT the question about major themes for 2023, the response was “I’m sorry, but I don’t have information on events or themes that occurred in 2023 or any time beyond my last update in January 2022. For the latest information on major themes of 2023, I recommend checking the most recent news sources or other up-to-date references.”
While not particularly helpful, I do find ChatGPT’s humility endearing.
To delve further into AI and what it might mean for the social sector, Provisio invited leaders in the Health and Human Services space to watch the Salesforce keynote (I bet you could guess what the focus was this year, even if our friend ChatGPT couldn’t.) We then transitioned into roundtable discussions with several thought-provoking questions.
AI in Human Services
One of the biggest discussion points was “where does AI fit in Human Services and working with people?”
Questions around AI include where the data goes, who shares it, and what are the ethical ways to use it. For example, we know that AI can help with automation, removing mundane tasks, but one participant expressed concern that there needs to be a place for everyone in their organization and would not want to see entry-level positions eliminated.
Another nonprofit offers digital literacy training for their clients and wonders about their responsibility for educating them on AI. How do they teach them to use it responsibly and what are the benefits?
Dr. Craig Maki, Provisio’s Chief Strategy Officer and former social worker, led the discussion. “It’s okay not to be okay if you are feeling confused or concerned,” he said, “because we’re all working on the question of AI together.”
Our position at Provisio is that technology and tools like AI will enhance the work you do. While systems can remove mundane tasks from your overall processes, it cannot replace the value that human beings bring to your organization, such as the ability to build connections with the people you serve and your community.
3 Principles for Trusted Generative AI
Perhaps you are wondering about the safety of AI. At Salesforce’s Dreamforce conference this year, one of the sessions covered their principles for trusted AI. You can watch the session yourself here, but here’s a quick summary.
Salesforce’s roadmap includes new features that enhance the accuracy of AI, like citing where the model pulled information from, highlighting areas where there’s uncertainty, and adding guardrails to prevent some tasks from being fully automated.
Salesforce makes every effort to mitigate bias and harmful outputs by identifying where the model delivers undesirable content. The Einstein Trust Layer helps keep data safe through safeguards like masking PII data in prompts and implementing zero data retention, meaning data is not stored unnecessarily beyond its immediate purpose. Data is never stored outside of SF or trained third-party models.
Salesforce won’t use any client data without their consent and clearly identifies when content has been generated by an AI model.
How to Prepare for AI
If your organization, like many we spoke to, is just dipping their toes into AI, there are a few steps you can take to prepare.
- Collect your data in one place.
- Ensure your data is up-to-date.
- Clean up your data. This means fixing or removing incorrect, corrupted, incorrectly formatted, duplicate, or incomplete data. AI results are only as good as the data it is given.
If you need help starting your AI journey, our team of experts can help you connect your data sources, migrate and cleanse your data. Contact us to get started
Marjorie lives in Saint Louis, Missouri. She received her Master’s degree in English from Truman State University. Her responsibilities at Provisio Partners include content creation and technical writing.