Business, Education, Finance, Government, Local, News, Technology

Ventura County Civic Alliance Livable Communities Newsletter – Nov. 15, 2024

by Community Contributor • November 14, 2024 • Comments Off

Volume 19 / Number 72 / August 2024

Your Livable Communities Newsletter

This quarter’s Livable Communities Newsletter reports on the 2024 FUTURE OF THE REGION CONFERENCE: HOW WILL ARTIFICIAL INTELLIGENCE SHAPE OUR REGION?. Our keynote speaker gave a balanced view of the positives and negatives of Artificial Intelligence (AI), and our panel discussed these points. However, this is not your typical conference report, as writer Kerry Roscoe soon found out. In interfacing with AI for these articles, she learned that AI can be a great partner in the quest to understand what AI is, and what it can do to shape our region. Read through the following three articles to see what surprises Kerry and AI have for all of us.

Let us know what you think about this.

Thanks,

Stacy Roscoe

Report on our Future of the Region Conference

HOW WILL ARTIFICIAL INTELLIGENCE SHAPE OUR REGION?

Keynote Speaker Jeff Butler:

Separating the Hype from the Reality

We have all heard the hype and the ensuing concerns. AI can do everything better and faster. AI will revolutionize the workplace. AI robots will replace humans. AI will learn to think and take over.

The reality is that, while AI can be an incredibly useful tool, it has limitations and the human element will always be needed to program, maintain, repair, and regulate it.

In terms of productivity, positive programs like Otter for notetaking are efficient, but AI can impact us in a negative way. Notifications per hour can be overwhelming and stressful. We have the “right to disconnect.” While it is not proven to be a cause, the overuse of AI has been shown to have a correlation with students who have ADD.

Efficiency is a big part of productivity. As an example, a weed killing machine created and run by AI was incredibly effective in using lasers for weed eradication on farmland, but at 1 MPH it was too slow to be useful. Economics plays a part in manufacturing, where the cost of the equipment and training is often too high to make it cost effective.

AI has been helpful in advances in medical technology, with things like the neural implant to assist paraplegics. It has been used to monitor risk in retirement homes and alert staff. While the Pero Robot dog is a new approach to therapy, at $6,000 it is too costly for widespread use.

AI has accuracy issues. It is not as reliable as we think. Many AI programs can “hallucinate” where some of what is put forth doesn’t match reality. Control and correction management by a human component is always necessary. Image recognition has shown to be accurate only 85% of the time. Things and Ideas with lower complexity are easier to trust. Complexity requires a person to monitor and check in order to inspire trust.

The reality is that AI can get it wrong. The old adage “garbage in, garbage out” really applies here. AI is only as good as its input. Microsoft’s Tay Bot using Twitter for data was a unmitigated disaster, creating content that was socially offensive and full of misinformation. AI uses existing data so it looks backward, not forward. We need to expand that knowledge base by including broader perspectives and information We need to avoid unconscious bias by recruiting, training and cultivating a diverse group of coders and programmers from across the globe. The more expansive and varied the input experiences and data, the better the results.

Social acceptance will play a big role in the future of AI. As initial excitement subsides, a more balanced evaluation can occur as the technology matures making both potential and problems more evident. For instance, we’ve seen pushback with autonomous vehicles and drones.

Both social conscience and economics will impact acceptance. AI requires a massive amount of energy to operate and a tremendous amount of water for cooling. There is both a monetary and an environmental cost. We may be able to do many things more economically with humans than with AI.

The bottom line: AI needs humAnIty to get it right.

AI Panel Discussion

Our panelists brought their experiences to give first-hand insight into how AI is already affecting us and how it could affect us in the future. As a lawyer who worked in the tech industry and as a professor of law, Jeanne Eicks was able to offer perspective in all three areas. Adam Lisagor, whose media company involves significant use of technology and serves many tech companies, shared both the benefits and downsides of tech in business. Dana Thompson’s long career in education and her position now as Director of Educational Technology and Information Systems at Oak Park Unified School District enabled her to provide insights into how education must evolve to address the changing needs of businesses.

Some Takeaways

As we look to educating the next generation to be our workforce, we need to acknowledge the role AI will play. What we need to teach the children is how to use AI responsibly and ethically: to know when and how they should use it requires critical thinking.

Teaching needs to look ahead to what skills our students will need in the future. AI can be used as a tool to customize curriculum for a student’s specific needs, but teachers will be needed to monitor and adjust for the human element. Use AI as a tool rather than a replacement.

Dana Thompson also noted that employers in the county have shared that they can teach interns their systems, and that what they need is the ability to communicate, write a cogent email, think critically, and know how to find resources for help.

On the legal front, Jeanne Eicks explained that students are already using AI to create study guides and test questions to quiz themselves for exams. The profession has long relied on technology for recidivism statistics (the rate at which criminals reoffend) to use in sentencing. On the down side, in trial prep AI has been known to cite cases that don’t even exist.

Additionally, there are many questions to be explored about intellectual property, content liability, and unintended consequences.

Information coming out is only as good as information put in. Data looks to the past, which can mean unconscious bias. For this to be truly useful and representative there needs to have e diversification on the input side and the inclusion of worldwide perspectives.

Adam Lisagor shared how AI programs like Summly can be used to summarize and give feedback that will push you to new thoughts. He created a program that summarized our keynote speaker’s talk. This provided interesting insight into the process which will be discussed in the last article of this newsletter.

Adam also emphasized that there always needs to be a human in the equation to monitor and check the system to verify its output. For instance, even if the weeding machine Jeff Butler highlighted in his keynote could be made efficient, it would still need trained people to program, maintain and repair it.

Another topic of interest during the panel discussion:

How do we stay safe and maintain privacy in an AI world?

An AI world can easily become overwhelming.

Be careful not to take on too much; self-regulate usage.

Trust but verify; ask the opinion of someone you trust.

Tip: If an email, communication or phone call generates an emotional response or a sense of urgency to act, PAUSE and take a breath. Then check/verify with another trusted source.

A Personal Take on the Effectiveness of AI in Notetaking

by Kerry Roscoe

To support my self-education, panelist Adam Lisagor graciously agreed to share the result from the AI program he had set up to summarize our keynote speaker’s presentation. I was curious to see how it compared with the notes I took, which were used for the first article in this newsletter. It seemed only fitting, after an AI Conference, to look deeper.

The AI summary appeared to me to be an excellent general overview of the areas covered, making it fine as a summary, but it was just that: an overview. It didn’t cover any of the speaker’s viewpoints or conclusions he might have had in those areas. What surprised me most was that one of the often repeated points in the talk was that there would always need to be a human component for oversight, etc. to keep it on track, and this wasn’t emphasized in this summary. The Final Thoughts section of the summary did reference that, but considering how much it was emphasized in the presentation, this seemed light. I wondered if there was a way to program the summary request to identify the speaker’s positions on issues, rather than just track the main topic areas.

When I spoke with Adam about this, he indicated that because it uses the transcript as its source, a tool can be configured to let the user ask questions and get finer details from it. When the transcript (think James Joyce & stream of consciousness with no punctuation) that Adam recorded with his AI system was dropped in ChatGPT, I was able to do just that. This proved most enlightening.

ChatGPT does not always answer a question in the same way. One summary of the talk actually enumerated the six focus points, while another answer did not, and I had to ask specifically..Also, Adam and I got different (but good) responses from the same platform to the question “What did Jeff Butler talk About?”

Interesting note: the ChatGPT self-corrected the transcript (yes, I read it all) when it answered some of my questions. It corrected the names of several authors and tech leaders. It also correctly adjusted “the father of Cider Park” to “the father of cyber-punk” and, when referencing the Metaverse, “having a poem strapped to your head” to “having a phone strapped…” among other good catches.

On the human side, my personal interests would have me emphasize certain things while AI pointed out things I overlooked. My memory had Jeff Butler discussing the high cost of electricity and water for AI in his keynote, when in reality he talked about water costs in the panel discussion. AI caught that for me.

My conclusion: Too err is human…, but that makes what we create unique.