A Busy Week of AI Exposures

            By Tuesday evening of a recent week, I’d had five  known exposures to AI. I’m intrigued, suspicious, and a bit horrified and it’s only Tuesday!

            The first exposure was participating in a study by a doctoral student for which I will be compensated with a small Amazon gift card. The study was described as:

“We are studying how to design mobile, wearable, voice-based, and traditional computing interactions to provide more accessible, efficient and enjoyable digital experiences.” Specifically, I sent them an image and then used an app they’ve designed to manipulate the image. I sent them a picture I’d Facebooked of a blooming amaryllis sitting on my dining table. I could get the image described, get questions answered about the image and then edit the image. Things I could do included cropping the image, brightening or darkening it and, adding text and stickers to it. After each edit it would describe it again. It was easy to use and definitely provided more possibilities for a blind person to work with an image than anything on the market now, that I know of.  Where it fell short in my opinion was it couldn’t make value judgments. For example, does adding a smiley face sticker to the picture add something or just look cheesy?

Next, I received an ad for:

Introduce +1 855-SHOP-GPT

Have you ever wanted a digital experience to feel like an in-person conversation? Innosearch AI is excited to announce our new, 100% free “shop by phone” experience to enhance accessibility by 10x. Shop using only your voice via your phone line. Browse billions of products, search from over 500,000 retailers, and compare products—all through simple dictation!

One has to sign up online using an email or social media. The service is great if you know what you want specifically, for example I asked it for samosas and it gave me reasonable choices at several stores. But if I wanted a new couch, it couldn’t tell me if it matched with the rest of my furniture. A person would have to be able to sign up online before being able to shop entirely by phone.

            Then came an email from the county telling me of a requirement to complete quarterly cyber security training because I’m a member of the county board. This quarter’s ten-minute training was about avoiding deep fakes. Bottom line, be suspicious about everything and verify requests for significant actions by some other trusted channel of communication before acting on them. Even if your boss asks you to send him a $5,000, check with him personally before doing it.

 Another email offered Gloo AI  which is a search engine with Christian values. It goes beyond just finding particular Bible verses to giving advice on “life’s big questions.” The devil is in the details, so to speak. Whose values?

            I’m all in when it comes to getting facts from Alexa, Siri, Chat GPT, etc. but I get queasy when it makes value judgments for me.

            To conclude my explorations, I had Chat GPT summarize this article, which it did a credible job of. It concluded with:

“In summary, the author is both intrigued and wary of AI, appreciating its utility but feeling uneasy about AI systems making value judgments for users.” Got that right!

.