Stephanie Nelson’s Post

Wondering how #AI is impacting the lives of #teenagers these days? Our new blog post by Thu Thu Naing dives deep into #generativeAI and its impact on teens! Check it out and share your thoughts! https://lnkd.in/gxEBNPR5

Is Grammarly Showing AI Responsibility or Are They Now Complicit in Widening the Equity Gap for Neurodiverse people? My work demands an incredibly fast pace of innovative ideas. So, I’m fortunate to have an ADHD brain that thrives in this environment. The constant cognitive stimulation and daily deadlines act like a dopamine Pez dispenser, letting new ideas, connections, and insights flow freely. The challenge? Getting these ideas out of my head in a way that's clear, organized, and understandable for neurotypicals can sometimes feel like torture. That’s where AI comes in. Since training my own GPT, I can pour my thoughts out as they come—messy, fast, and all over the place—and my GPT helps me organize them, flagging information gaps and areas of potential confusion, then structuring things in a coherent way. Sure, I still need to work with it a bit, edit it and throw in some personality, but in the end I've saved hours, sometimes days. AI doesn’t think for me; it works with me, allowing me to bring my full potential to the table. For the first time in my life, I have an assistant that can keep up with my brain’s pace. 🎯 But today’s Inc. Magazine article about Grammarly's new Authorship tool introduces a troubling narrative. For context, Grammarly Authorship aims to track whether AI wrote or helped edit a document. While transparency is important, this tool risks widening the equity gap for neurodiverse people who rely on AI to communicate our ideas. Those with Dyslexia or Dysgraphia who struggle with reading/writing, or ADHD and Autistic people who wrestle with organizing their thoughts, often find text-heavy communication limiting at best. AI is an incredible asset to help them express their genius and share their voices which have been missing from most future-shaping conversations in human history. Look, I understand students need to learn how to research and write papers, and if teachers want to use it to make sure the student did their own work, fine. But mark my words, writers who feel threatened are going to start throwing mud and canceling people. Coworkers who are afraid of becoming obsolete are going to start shaming their AI using colleagues and having tantrums so big even Louis Litt will heel turn in the other direction. AI is here, and it’s not going anywhere. So, as we continue discussing how it should or shouldn’t be used, let’s make sure we don’t rest until we’ve contributed to a more inclusive and empowering society—celebrating rather than shaming each other for having the audacity to expedite innovative ideas and amplifying the voices previously excluded. For now, I'm going to have faith in Grammarly and their AI Responsibility practices, trusting they understand their place in this part of history. We are all shaping the future, let’s give it a chance to evolve from our best intentions and our brightest dreams, not from our fears and biases. 🌍✨ #Neurodiversity #ADHD #AI #Innovation #Inclusion #ChatGPT

Generative AI is transforming child language learning. Researchers have crafted a cutting-edge personalized storybook generation system powered by AI to enhance children's language skills. This innovative system, recently honored at ACM CHI, adapts vocabulary learning to each child's unique requirements and surroundings. By analyzing language patterns and crafting tailored storybooks, the system offers a more efficient approach to language development. Read it here: https://lnkd.in/etnigaFp #AI #Education #LanguageLearning

I finally got my #AI field-testing paper published! Open access! It's the 1st paper that I know of that explores using AI examinees to field-test new test items. I improved the proposed method since the last version of the paper. Thank you Smarter Balanced for the support! I will continue research in this area #psychometrics #research #irt #bert #LLM #education #assessment

🧠🌐 Embark on a journey through the fascinating realm of neuroplasticity and second language learning with Li, Legault, and Litcofsky’s groundbreaking study. Published in Cortex, this research illuminates the remarkable anatomical changes that occur in the human brain as a result of second language acquisition. Li and colleagues provide compelling evidence of the brain’s capacity for adaptation and reorganization in response to language learning, highlighting the transformative impact of bilingualism on neural structure and function. 📚💡 Through meticulous analysis of neuroimaging data and experimental findings, Li et al. offer valuable insights into the neural mechanisms underlying second language acquisition. Their study reveals how learning and using a second language can lead to structural changes in key brain regions associated with language processing, memory, and executive function. By documenting the plasticity of the brain in response to linguistic experiences, this research underscores the cognitive benefits of bilingualism and the potential for language learning to enhance cognitive flexibility and resilience across the lifespan. 🌟🎓 Let’s celebrate the remarkable plasticity of the human brain and the transformative power of second language learning! Inspired by Li, Legault, and Litcofsky’s research, let’s embrace bilingualism as a pathway to cognitive enrichment and neural resilience. By promoting multilingual education and fostering a culture of linguistic diversity, we can unlock the full potential of the human brain and cultivate global citizens equipped to thrive in an interconnected world. #Neuroplasticity #LanguageLearning #Bilingualism #ScientificThursdays Cite this article: Li, P., Legault, J., & Litcofsky, K. A. (2014). Neuroplasticity as a function of second language learning: anatomical changes in the human brain. Cortex; a journal devoted to the study of the nervous system and behavior, 58, 301–324. https://lnkd.in/dHQPsYxW Find out our guides and resources specially handcrafted for you fellow educator. Link in our bio. #4Learnings #learninghowtobetterlearn

🤖 Check out how companies are using Meta Llama – our open-source large language AI models – to shake things up in education, customer service, research, and even medicine! 🚀 From tailored math learning platforms to AI companions for your Zoom meetings and providing medical info where it's needed most, Meta Llama is making a difference. With over 170 million downloads, Llama 2 and 3 are rocking the AI world! 🤩 #MetaLlama #AI #Impact #opensource

I have a unique perspective on the usage of Large Language Models: I am Autistic. I struggled so much to understand and predict my allistic family and peers that I ended up getting a Bachelor's Degree in English. (I basically reverse engineer the internal states of everyone I interact with based on their word choice, cadence, and sentence structure. I do pick up on the existence of strong emotions, but can't always read which emotion it is.) I am teaching my neurodivergent, low-masking nephew the differences between exclamatory, imperative, interrogative, and declarative sentences, since he has been struggling with the explanations he has gotten in class. (We'll be working on quotation marks and dialogue tags next.) I spent a lot of time with the researchers in the AI Lab at the University of Wyoming when I was attending college. I got to have firsthand conversations about what the greater goal of training neural networks was for each of the individual researchers. I may be over-interpreting their reasons, but the pattern seemed to be that they wanted to uncover the truth of existence through emulating creation, recognition, imitation, and evolution, across species. I've been on the implementation side of LLMs and this obsession with shoehorning it into everything (which may be helpful for allistic people. I don't know. I do know that I'd rather learn entire subjects before speaking to a human when I don't absolutely have to.) Here is my conclusion/opinion: Human communication is most efficient at conveying in-group or out-group status, not at conveying information itself (think memes, shared alma mater, trauma bonding). As humans, we seem to put a lot of weight on that in-group/out-group status as a matter of survival, not higher pursuits (even though we pretend otherwise.) Therefore, the injection of LLMs into customer service and advertising is meant to manipulate that survival mechanism into eliciting unearned trust from an increasingly savvy consumer. (It is difficult to earn real trust with bad products and business practices.) I'm curious about what this may lead to when it comes to human communication in the long term; if we will become more divided, or if this will be a commonality that binds more of us together. I suppose there is not much more for me to do other than wait and see.

Raising Kids, Theory of Mind & Artificial Intelligence Theory of Mind (TOM) typically develops in children around the age of 4-5, when they begin to grasp that other people can hold beliefs different from reality or their own knowledge. It is a fundamental aspect of human cognition, allowing us to understand others' thoughts, beliefs, desires, and intentions by deduction. It is also defined as "the ability to impute unobservable mental states to others". It’s what helps us navigate complex social interactions, predict behaviors, and build strong relationships. One of the most famous tests of ToM is the False Belief Task, developed by psychologists Heinz Wimmer and Josef Perner in 1983. In this task, children are asked where a character will look for an object that was moved without their knowledge. The ability to answer correctly demonstrates an understanding of others' mental states—a key milestone in cognitive development. While AI systems excel at tasks like pattern recognition, data analysis, and natural language processing, it would be a big leap for LLMs to demonstrate Theory of Mind. Why does this matter? Imagine an AI that can: ✅ Accurately predict human emotional responses and intentions ✅ Adapt its behavior based on understanding a user’s beliefs and desires ✅ Collaborate more seamlessly in human-machine teams by anticipating human needs Now check this linked Stanford article. Experiments already showed that GPT-4 demonstrated the capability that is typical of a 7-year old human (i.e., accurate on 95% of false belief tests). This is claimed to have 'spontaneously emerged'. Besides being a fascinating step towards deeper AI-human interaction, the fact that this spontaneously emerged indicates to me the importance of language in learning. LLMs are inherently capable, but LEARN through natural language communication. That response-stimuli feedback loop is known to be so important to raising kids, and clearly is an important step towards the development of LLMs. Creating machines that can understand and adapt to human mental states is can (and is already beginning to) revolutionize fields like healthcare, customer service, and social connections. Big power, big responsibility. https://lnkd.in/dWTV5C-q

🔍 AI: Learning Like a Baby 🍼 🤖 Revolutionary AI Study: - A groundbreaking AI model has mimicked the learning process of an infant, offering new perspectives on human learning. - By analyzing just a sliver of a baby's experiences through headcam footage, this AI learned words like "crib" and "ball" without prior language knowledge. 🧠 Challenging Traditional Theories: - Contrary to some cognitive science beliefs, this study suggests babies may not need innate language knowledge to understand words. - This AI's learning process is based solely on associating images with spoken words, pushing the boundaries of our understanding of language acquisition. 👶 Unique Baby-Eye View Approach: - Researchers used 61 hours of headcam footage from Sam, a baby in Australia, capturing his world and interactions. - This approach allowed the AI to learn from an authentic infant perspective, a novel method in the field. 📈 Impressive Learning Outcomes: - The AI's ability to match words with images reached a 62% success rate, showcasing significant learning from minimal data. - It even recognized objects it had never seen before, a testament to its learning capabilities. 🔬 Challenging Established Linguistic Theories: - Findings challenge the idea that language learning requires special mechanisms, supporting the potential for general learning processes in early language acquisition. 🚀 Future of AI and Cognitive Science: The study opens up exciting avenues for further research into AI's potential to mimic and understand human learning processes more deeply. 💡 Takeaway: This innovative research not only sheds light on how infants learn but also paves the way for advancements in AI, potentially transforming our approach to understanding human cognition and language development. 🌟 Stay tuned for more updates on this fascinating intersection of AI and cognitive science! Credit: Nature - full article in in first comment

One focus of CBRs research is to understand #aphasia in #bilingual #people with #stroke and we use #AI and #computational models to #simulate and #predict language behavior https://lnkd.in/eNUnfGhU