Chat Generative Pre-trained Transformer, or ChatGPT, is the hype term of the moment. The interest comes from ChatGPT abilities to create content and provide conversational results to an inquiry. It essentially promises to connect artificial intelligent processing with automatic conversational responses.
The applications for customer service are obvious. Also perhaps obvious are applications that could augment or replace “search,” or “writing.” As TikTok alarmed Facebook, perhaps ChatGPT now alarms Google.
The big deal is the ability to create content based on existing content and data. It is not so much a use case related to the equivalent of human “thinking” as to “content creation” based on precedent and existing data.
Generative AI is the larger trend that ChatGPT is part of. AI-created original content is the promise. An annual or quarterly report, for example, is a fairly structured document drawn from existing data, the sort of thing generative AI is supposed to be good at. News stories, sports scores and advice (legal, financial, business strategy, for example) are the sorts of content that are based on existing formats, precedents, databases and conventional wisdom or rules of thumb.
When to buy a product; why it provides value; how to buy; where to buy; from whom to buy; understanding pros and cons are some of the questions generative AI is ultimately expected to provide.
What options might be in any legal matter, what the precedents are, and what courses of action can be taken are legal questions all based on past experience. “How to invest, at a given age, with assets of different amounts, with defined goals, in what instruments, for how long, and why” are all questions with answers based on clear rules of thumb used by financial advisors.
The uses in education, which mostly consists of knowledge transfer, are endless.
It probably is not too hard to see how generative AI could be used to create personalized marketing, social media, and technical sales content (including text, images, and video).
Some believe generative AI could write, document and review code. Applications in many other fields, ranging from pharmaceutical development to health outcomes, in fact all human endeavors with large existing data sets and “expert” advice, could be enhanced.
Anywhere there are patterns in data, and lots of data to be worked with, it is possible that generative AI could add value. The more complicated processes are--such as weather--the more value could be obtained, in principle. Generative AI essentially creates based on existing data. So the more data, the more creation is possible.
Is that “new” content derivative? Yes.It is based on the existing data, which can be manipulated and displayed in original ways. And generative AI is about creating content, not “thought.” But content creation is expensive and important in almost every sphere of life.
The hype will pass. But disruption and substitution clearly can be seen as possible outcomes, eventually. Anywhere content has to be created, where there are existing rules of thumb about what is important, where lots of precedent and data exists, where some questions have obvious standard answers, generative AI is likely to be valuable and important.
It is not simply content creators, but advice givers that could ultimately see their output devalued. If you ask me when MPLS adds value, and why, and how it compares to SD-WAN, there are a limited set of answers I can provide that correspond with industry wisdom about such choices.
Over time, a greater number of questions will have answers computers can assemble and deliver. It’s coming, if not right away.
No comments:
Post a Comment