One way people try to appeal to me in regards to artificial intelligence (AI) is by telling me it will make data more accessible. As the API Evangelist I am regularly advocating for companies, organizations, institutions, and government agencies to make their data more accessible using HTTP APIs. Which is a very nuanced argument which people misread and misinterpret in some pretty telling ways. The people who are believers in AI know this about me, and regularly try to appeal to my desire to just make data more accessible, without actually having read my work, understanding my position, or seeing the finer details of API management and service composition and how this plays out in making data “accessible” via the Internet.
Not all, but a significant portion of the AI discussion echoes previous open data discussion, so it makes sense that the argument continues. It took me 2-3 years to figure out that a significant portion of the people I was involved with in the fight for opening up data weren’t actually on my side-—they were just using me. 2010 through 2015 there was a huge cry for companies, organizations, institutions, and government agencies to open up their data and make it freely available for download or via APIs. It is an argument that sounds altruistic, but once you realize the majority of people calling for this have no interest in opening up their data, they simply want to convince you to open up yours. Open data actually meant open for business using your data and other people’s digital assets.
This phase of artificial intelligence is just the future of the open data movement. AI doesn’t make data more accessible to those who don’t have access to it. That is a frontline illusion to convince you personally or the company you work for to open up your data and other systems so that you can be plugged into this new magical realm. While there are 5-15% real world use cases where AI can make data more accessible to people this does not make up for the larger percentage of use cases that purely want to extract your data and produce value for the people producing the large language models. The disrespect for copyright in the push for LLM demonstrates these real-world motivations behind AI making your data more accessible, and that the people producing LLMs aren’t interested in your access to data—-it is about their access to your data.
I have spent fifteen years watching APIs begin with a promise of access and democratization only to watch good API after good API go bad and become extractive and exploitative. I am sorry, you aren’t going to convince me that AI has any other motivations other than extracting value and exploiting those who put it to work in their personal and professional lives. I spent way too much time with my ass hanging out there making the same argument for APIs, while the leaders and investors of whatever API, service, or tool I was advocating for walked away with big paychecks. If you really want to have a conversation with me about how Ai makes data more accessible feel free to come by for an API Evangelist Conversation and we can test out those assumptions and beliefs. My guess is you are either comfortably in denial about how AI Is being used because you perceive yourself as above the AI line, or you are in on the hustle—-either way, I look forward to learning more from you about how AI makes data more accessible to the common person.
