On The Road With Al & Ivy: June 16, 2026 Riffing On Street Level AI: The DIY aspect Note: The views expressed here are opinions and are one person’s ruminations on the subject. Also, I’m making it a point to not make the text too polished in this new AI era. One of the biggest achievements by the AI industry is making most people regard an AI bot as a living thing and thus to blame for any stolen creativity or social harm. It’s like the notion of sin being caused by the devil, or some evil influence as opposed to the human impulses that create unethical behavior or crime.  The principle is that you can’t do anything about a devil, but you can definitely do something to the disciple; either to eliminate the heresy or punish sin. That creates the irony (and distraction) of luddites and enemies of Big Brother AI protesting on the Internet (which is rapidly becoming fully powered by AI and run by management who support it) and calling for boycotts or public shaming of AI users, regarding it as behavior as bad as, say, looking at your phone during a public event or knowingly consuming white sugar. I believe the standard conspiracy term for such a phenomenon is “divide and conquer.” There’s no tech history that would indicate that street level artificial intelligence, in spite of the hype, will become any more proficient than autocorrect or robot customer service; both are technologies that have been around for decades but still make mistakes.  I’m talking about the level of service the general population gets, not the performance for those who pay a premium. One definite reason for the institutionalized mediocrity of AI is because the industry didn't quite pull off the scale of data scraping that it was hoping to do. I imagine that the firms were taken by surprise by the scale of resistance to scraping the Internet (or the building of the data centers necessary to power the software).  However, the human process of information consumption (ideally) involves using public domain data for free and paying in some way for information that’s copyrighted. Also, works that involve research generally include credits for source material. In other words, there’s a cooperative and legal process for data use that the AI industry ignored in the early stages. …your results may vary… Once AI firms had to start licensing data and major sites and social networks started demanding their piece of the pie, then the technology, which has to be worked on and refined, will always be constrained by cost and access to data. In other words, the quality of the AI is going to vary, depending on how much the company is willing to pay for the necessary data, or cares to. …some solutions… However, never underestimate the ingenuity of the high-tech industry, or capitalism in general. One solution is to sell or license the AI software to a middleman, who then collects the data by means that are already acceptable to the public, i.e. giving up their private data to use social media sites or creative apps for free. Another way is the traditional software method of making the user train the AI, or help do it. Like with autocorrect, which requires a user to help teach the software or customer service bots which are labeled as experimental (and all of your interactions with it are helping to train the robot for free). If consumer cooperation stops, there’s no software code that can replace the date provided. It’s very much dependent on data, not algorithms. One thing that’s clear is that AI is definitely here to stay because it automates a great number of tasks that would normally take some money or time by an average person. AI is already winning the war by numbers. In the next blog entry, I’ll riff on the reasons why street level AI is growing so fast and the past historical examples that clearly show the effect the new tech will have on jobs. Al Handa May 2026 #AI #artificialintelligence #blog #hightech #software