Carlos Romero’s Post

If you have your doubts about AI, you’re in good company: 58% of Americans lack faith that AI can aid in moral reasoning, according to our new #StateoftheBible report. Discover more about the intersection of faith and AI by downloading our free report today: StateoftheBible.org

"Bear in mind that our world is dominated far more by human stupidity than by AI." Are GenAI concerns all fear-mongering, when humans can be quite 'stupid'? This raises an important point - humans are the designers and algorithm generators. If there is human stupidity, should we be concerned about the extent to which GenAI is being used? What do you think? Share your thoughts in the comments below. Article (in the comments) for more insights. #dataandanalytics #dataengineer #genai

Ever thought AI might just be showing us our true selves? In this article, written by our Creative Thinking in Residence, Sophie McKay Knight, we unpack how AI mirrors the best and the worst of us, offering a unique glimpse into human nature through technology. Intrigued by what AI reveals about us? Discover the insights here 👉 https://bit.ly/42I4lNd #AI #Reflections #TechExploration #ThinkingZone #DigitalLearning #LearningTechnologies

Have we reached Peak Human? This piece ( https://lnkd.in/g6mFuNtd ) is an expanded version of my VentureBeat article from last week in which I explore this issue. By "Peak Human" I mean the moment in time when AI systems are just as smart as half the humans on planet earth. After that, we will be on a steady decline until eventually all humans, even geniuses, are outmatched by AI. As described in the piece, I believe we are not far from Peak Human by many measures, including basic logic and reasoning. This surprises me, but I can't deny that Open AI's new o1 model does shockingly well on reasoning tests, including IQ tests it has never seen before. Yes, it still makes dumb mistakes at times, but so do most humans. Probably all humans. More details here: https://lnkd.in/g6mFuNtd Unanimous AI David Baltaxe Joshua Sitzer

Enhancing human creativity, rather than replacing it, and providing time for creative pursuits, are appropriate ways to view the relationship between AI and human creativity. Melissa Wheeler captures this well when she writes: “By freeing us from routine tasks, AI allows us to dedicate our time and energy to what truly inspires us. As we continue to embrace AI, let’s ensure that it remains a tool that empowers human creativity and enriches our lives. In this symbiotic relationship between humans and AI, the future holds endless possibilities for creativity, innovation, and a more fulfilling life.” #humanskills #creativity #AI https://lnkd.in/e-K_2aRM

Hallucinations and LLM ** Next in the series of a CEOs AI agenda ** All LLM models are probabilistic models. So, by definition, they would produce results that with a "Non-zero" probability will be senseless to a human. It's a feature and not a bug. We must improve our ability to handle the output that does not make sense. Trying to make hallucinations go away is a losing battle. #ai #ceoinsights #LLM #hallucinations

How much effort to get rid of the text errors in this AI-generated image which at first glance seems production-ready? Humans would never make such mistakes, and no AI can find them or fix them, so how much human work is needed to fix just one nearly-perfect image? #ai #genai #generativeai #aiimagegeneration #hype #errors #humanintheloop

Louis Rosenberg suggests we're reaching peak human. That the AIs are on the verge of being more intelligent. I've said this a few times now: The knowledge metric for intelligence is entirely the wrong way to measure intelligence in machines. It's judging the intelligence of a dictionary based on its vocabulary. Of course, AI has a lot of knowledge; you had to stuff the entirety of human knowledge into it to get it to work at all. Moreover, virtually all of these systems repeatedly show that their intelligence is wildly inconsistent. If you ask a human what 2+2 is, and they get it wrong 25% of the time, then you're going to have to do more testing to see why, and the assumption shouldn't be high intelligence. Louis mentions this in his article, but says " we see evidence that AI systems are catching up with us just as quickly. It was only a few years ago that virtually all artwork created on planet earth was crafted by us humans".  You are saying that what the AI is creating is art.  I would also have to disagree with that.  It's a distillation of human art, not actual art of its own.  The AI has no perspective and is incapable of imbuing whatever thing it's created with emotion.  The only reason it can do anything at all is because it's interpolating the human perspectives it has ingested. I've yet to see AI have any ability to create something novel; at best, it's a statistical attempt to meet the creativity in the prompt. When he places collective human intelligence on the graph, it makes me think of how AI is a stupider version of human collective intelligence.  It took several hundred million people to be as intelligent as 35 average people. Which is kind of sad. #ai #artificialintelligence #llm #art #creativity #agi #openai #google #microsoft #meta #anthropic

AI models/algorithms are dependent on what we, as humans, design. AI performance is dependent on its feeding data. AI (ChetGPT x/OpenAI o1) feeding data almost finished scraping from the Internet. I wonder if AI could evolve more if no more human data is available. If the statement above is true, confining human data input into the AI is a way to protect us from future AI harm. CASS is a good solution. Please refer to US10972256 or www.Archss.com.