Are you tired of the hype surrounding AI? Do you wish someone would just cut to the chase and say something useful about it? Well, here are 10 fresh thoughts that should help a lot since they’re drawn from my most recent interactions at conferences, in association chapter meetings, and with my clients.
Let me know what you think, and what other perspectives you have. Thanks!
- AI is math, not magic. It utilizes significant computing power to analyze all the information it has at its disposal and essentially calculate which possible result has the highest probability of meeting the user’s parameters.
- The decades-old principle of “Garbage In, Garbage Out” (GIGO) is very much in play.
- AI generally isn’t present where we think it is, especially in information applications. In fact, the majority of solutions claiming to be AI actually use machine learning, which is AI-“ish” and extraordinarily effective, but self-learning rather than self-teaching like “real” AI is.
- Practically speaking, this distinction is largely inconsequential. However, it is important to understand what you’re buying to avoid being oversold in terms of price and capability.
- AI is incredibly fast and impressively capable. It can process highly complex queries far more efficiently and effectively than the human brain can.
- This is the root of its general appeal – and also of its absurd demand for electricity.
- AI is just a tool. It can be any kind of tool you need it to be. Today, it most often is being used as a search engine, a ghost writer, a research assistant, a code generator.
- This versatility is quite attractive but also makes it hard to define AI or characterize what it is/does.
- AI often is not even necessary. Chances are you already have other software that can handle most of the needs identified as being AI-worthy (e.g., to search/find records, analyze data, process task requests, take meeting minutes) or can get it relatively cheaply.
- Microsoft 365, Zoom, and other similar services already include many of these capabilities. These may not be as fast or robust as a dedicated AI installation, but they will still let you get the job done without qualifying as overkill.
- AI returns results without regard to accuracy, timeliness, propriety, or other quality criterion. Results must be validated by human users who are qualified to evaluate their efficacy. “The computer said” is not a valid defense if/when a problem arises.
- Providing context as part of a prompt can mitigate this to a certain extent, but true discernment can only be provided by people.
- AI prompt-writing is an important part of ensuring quality results. An emerging new art, it may soon become a sought-after resume skill and perhaps even a new job function.
- Questions are now being raised as to whether/when AI prompts ought to be considered records.
- “Public” AI services (i.e., those you can use for free) do not maintain the privacy of any information you input. They use this information to improve their models and to serve as the basis of the responses provided to other users.
- This is true for the prompts you submit and any source material you provide to the service to improve the relevance of the responses you receive.
- AI is not going to take your job or rule the world. Its increasing visibility, however, will transform the way we work, not by direct application but by raising the service expectations of the residents we serve.
- People young and old are growing accustomed to receiving easier access to information they need and faster responses to the questions they ask. Though not a product of AI specifically, this expectation is adding pressure for us to perform with what may be seen as “AI speed.”
- AI can be used either for good or for evil. In this regard, it is no different than any other invention.
- Which way it goes ultimately is up to us.