You can’t listen to a radio or TV news program or read a newspaper without running across an article about some aspect of artificial intelligence (AI). Much of the debate centers around questions about losing jobs to AI, misinformation and deepfakes created by AI, bias and fairness, and who is responsible if AI makes a mistake or causes harm. Other debates focus on copyright and creativity issues. Certainly all legitimate concerns.

While these issues are being sorted out, the rest of us may be interested in what AI can do for us. Can it help us problem solve everyday issues? Is it really what some have called a “search engine on steroids”? Is it accurate enough to be useful? Is it really faster than a more typical internet search?

Generally, I’ve found that AI can be helpful for answering specific questions and solving problems. In these examples, I used my two go-to AI chatbots, Claude and Gemini, to help me figure things out.

Please note that I wasn’t using AI in these cases to write articles or otherwise create content … or to write a romantic card to my wife! I was simply trying to use AI to fix, solve and understand several personal tech problems.

Sometimes AI was very helpful, sometimes frustrating and, occasionally, it wasted my time and was just plain wrong! I hoped to problem solve these several tech issues without spending lots of time searching random websites or watching YouTube tutorials.

A couple of interesting tech issues popped up. It started with messages that were not syncing between my iPhone and Mac laptop, even though everything was set for them to sync. Then a mysterious orange light appeared in my Dynamic Island (the pill-shaped section at the top of my iPhone 16) that I could not turn off no matter what I did.

The next day I saw SOS at the top of my phone, something never seen before. Later I found out I had “participated” in Verizon’s nationwide outage — no iPhone for several hours.

Two home issues rounded out my week — working to connect a new garage door opener to my car’s garage door button and learning to fix the agitator in our washing machine that came loose.

A couple of examples of how I used AI

The orange light saga on my iPhone was a long one, and I had a lot of back-and-forth conversations with Claude about this one — 18 questions that I asked. And long lists of things for me to try to get rid of the orange light. As it worked out, through trial and error, I turned off Vocal Shortcuts in Accessibility and that turned off my orange light.

Claude replied, “Excellent detective work! I’m glad you found it — Vocal Shortcuts was the culprit all along. That makes sense now — it was keeping the microphone active in the background, which is why the orange indicator stayed on. Sorry for all the runaround with the MFi Hearing Devices. Thanks for sticking with it and figuring it out through trial and error. And thanks for letting me know what actually worked — that’s really helpful information!”

The washing machine fix was longer than it should have been for a relatively simple repair, but the difficult part was Gemini sending me fixes for other machines, not the one in my house.

What I learned

Did AI help me answer the tech questions I was dealing with? In the instance with Verizon’s outage that shut down my iPhone, Claude gave me lots of ways I could check to see what might be causing my phone not to work. On the other hand, Gemini immediately told me about the Verizon outage and pinpointed that as the issue. Problem solved!

With a couple of the other tech problems (the orange light on my phone and setting up the internal car button synced to my garage door opener), the AI responses gave me a lot of things to try but did not immediately help me solve the problem. Eventually, I gave up on the garage door opener sync. AI couldn’t help me get it to work.

How to get helpful responses from AI

Be specific and detailed. For the washing machine fix, I submitted several pictures of my agitator and that was exactly what AI needed to find the directions for my specific machine. Specific directions (model numbers, photos, etc.) are key.

Provide context. Let AI know what you’ve tried, what constraints you are working with and your end goal. For the orange light issue, it took me several questions where I provided enough information for AI to understand what I was asking. That was my fault for not addressing at the beginning.

Iterate. This may be the most important part. If the first AI response isn’t quite right or doesn’t solve the problem, tell AI what’s missing or what needs adjustment. AI will then refine their answers based on this feedback. I did this in each instance and it made a big difference.

Ask for step-by-step reasoning. This is especially helpful on more complex problems. Ask AI to show their reasoning. That was helpful on the orange light issue and helped me understand the possible solutions more than I would have otherwise.

Remember, the key difference from traditional search engines is the conversation — you can ask follow-up questions, request clarification or say “That’s not quite what I meant” until you get genuinely useful help rather than just links to wade through.

And also remember that your AI chatbot is far from infallible. Claude reminds us at the bottom of every page that “Claude is AI and can make mistakes. Please double-check responses.”

Using AI to help me problem solve several tech issues was really a mixed bag. Some things worked. Some AI explanations were off base or didn’t solve the problem. But using two different chatbots was helpful to keep Claude and Gemini honest.

BoomerTECH Adventures (boomertechadventures.com) helps boomers and older adults navigate the digital world with confidence and competence. Active boomers themselves, they use their backgrounds as Maine teachers to support individuals and groups with online courses through Maine Adult Education (tinyurl.com/BTAclasses) articles, videos and presentations to organizations upon request.

The Times Record Sustaining Sponsor
We believe a community must be informed to thrive. bowdoin.edu