JP Liang

We already know the transformation that the use of AI and especially GenAI implies. We must now work on transforming teaching methods and evaluation processes. I share this link from the University of Cambridge. I Select this : "Promote Process over Product: Encourage students to use LLMs as a supplement to their own work, not a replacement. Students could focus more on the processes involved in problem solving, project iteration or essay writing over the final product. The goal should be to enhance their learning, not bypass it. The inherent benefits and focus of a programme of study or discipline should be shared with students and promoted regularly" I return to this point Human Intelligence + AI = Augmented Intelligence, The big challenge is how we design it. #GenAI #AI #Education #AgI

14

AI—Existential Threat or Paradigm Shift? The recent Physics, AI, and the Future of Discovery fireside chat, moderated by Science Philanthropy Alliance President France Córdova and sponsored by the American Institute of Physics (AIP) Foundation, was a spirited discussion exploring the intersection of AI and physics. Key speakers included notable scientists from industry, academia, and government sectors, who shared insights on how AI is transforming physics research and its potential future implications. The discussion focused on the transformative potential of AI in physics and covered a range of topics, including: 📊 AI's role in theoretical and experimental physics, enhancing data analysis, and accelerating discoveries. 🤝 The synergy between AI and physics, where each field enriches the other. 💡 Industry perspectives on AI as a tool for innovation and productivity. 📚 The importance of AI in education and outreach, making complex concepts more accessible. 🔍 Ethical and policy considerations to ensure AI's reliability and trustworthiness. 🌐 The importance of collaboration between academia, industry, and policymakers to harness AI's full potential responsibly. Watch the full discussion here: https://lnkd.in/gigrajxA This event was hosted by Michael Moloney, CEO of the American Institute of Physics, and featured renowned experts Dr. Walter Copan (VP of Research & Technology Transfer at Colorado School of Mines), Jesse Thaler (Institute for Artificial Intelligence and Fundamental Interactions and Massachusetts Institute of Technology professor of Physics), Valerie Browning (VP Research and Technology at Lockheed Martin), and Evgeni Gousev (Sr. Director at Qualcomm AI Research).

14

1 Comment

Here is another study looking at the use of AI in education. This paper by Owen Henkel, Hannah Horne-Robinson, Nessie Kozhakhmetova, and Amanda Lee titled "Effective and Scalable Math Support: Experimental Evidence on the Impact of an AI- Math Tutor in Ghana," explores the potential of Rori, an AI-powered math tutor available on WhatsApp, to enhance math learning in resource-constrained settings.  The intervention provided extra math instruction to a treatment group via short weekly sessions with Rori, while a control group received standard classroom instruction without Rori. Results showed that students who used Rori made significantly greater progress in math, with an effect size of 0.36, equating to approximately a year of additional learning. Key Takeaways: Significant Learning Gains: The treatment group using Rori showed marked improvement in math performance, underscoring the efficacy of AI-powered tutoring in low-resource settings. Cost-Effective and Scalable: Rori operates on basic mobile devices with low data costs, making it accessible and financially viable at around $5 per student in low- and middle-income countries. Future Implications: The success of this chat-based, AI-driven approach indicates potential for scalable, personalized learning solutions globally, especially in under-resourced educational environments that lack traditional tutoring infrastructure.​

3

Panel Title: Realities and Potential on the Africa Continent - Stanford Institute for Human-Centered Artificial Intelligence (HAI) The panel suggests that there is a need for systemic changes to ensure that AI adoption in Africa benefits the continent's people and promotes economic growth, innovation, and development. #Speakers Justin Arenstein | CEO, Code for Africa Barbara Glover | Program Officer – African Union High-Level Panel on Emerging Technologies, African Union Development Agency-NEPAD Vukosi Marivate | UP ABSA Chair of Data Science, University of Pretoria Kiito Shilongo | Senior Tech Policy Fellow, Mozilla Some key points mentioned by the panel include: - The importance of language as a catalyst for AI adoption in Africa - The need for guardrails and policies that ensure responsible AI development and deployment - The challenges of coordination between different countries, organizations, and stakeholders - The potential risks of AI being used to perpetuate social and economic inequalities - The need for normative frameworks and guidelines to promote ethical AI practices - The importance of human rights in the development and application of AI https://lnkd.in/eHH3Wsft #AI #africa #afrokwary

2

The 2024 State of CS Report is out! Download the report, review the data, be inspired by the educators and states highlighted in the report, and ask yourself how you can best position yourself and your state to advocate for broadening participation in computing. We have work to do!

18

From Dr. Jeffrey Funk. Most entrepreneurs are not, in general, technologists. Some are, but the overwhelming majority of people who go into business out of an MBA program are usually far more interested in financial returns than in the technology itself. This often means that there is usually a huge amount of magical thinking that takes place in business, because these are NOT experts in the technology, and they are dependent not only upon their pet wizards (who usually are only secondarily interested in the monetization aspect) but also on their own understanding and perception about the technologies and the markets these technologies address. When such technology gets beyond the skill sets even of most "experts" in that space (which is where we are now), the ability to evaluate that technology and its potential becomes even more suspect. While I think it is still "early" in terms of the potential of generative AI in particular, there are several key holes that still need to be filled first before it becomes more than a moderately break-even proposition. These aren't being taught in business schools, and there's real danger there because of it.

10

📢 Exciting Announcement: Community Partnership! We’re thrilled to unveil a powerful new alliance between Melanated Mathematics®️ and Black AI Media™️—two visionaries in educational transformation and tech-driven empowerment. Together, we’re merging culturally relevant learning with groundbreaking AI innovation to deliver next-generation educational tools designed for true impact in Black and underserved communities. Why this Partnership Matters As champions of representation and equity, Melanated Mathematics®️ and Black AI Media™️ bring unique strengths to this collaboration: educational expertise and AI innovation tailored to the needs of our community. This partnership fuels our shared mission to empower Black students, families, and educators by making high-quality, culturally affirming STEM resources more accessible than ever. What to Expect 🔸 Engaging, AI-driven content: Think dynamic, interactive lessons that inspire curiosity and critical thinking. 🔸 Culturally relevant resources: Reflective of our students, our stories, and our legacy in every equation. 🔸 Community-first innovation: We’re prioritizing feedback and guidance from parents, educators, and community leaders. Together, Melanated Mathematics®️ and Black AI Media™️ are creating tools and opportunities for growth, so our students not only see themselves in the learning process but feel empowered to shape the future. Follow us on this journey as we bring meaningful, accessible, and culturally-rich educational experiences to the next generation. 🚀 👉🏾 https://blackaimedia.com 👉🏾 http://melanatedmath.com #MelanatedMathematics #BlackAIMedia #STEM #CulturallyResponsiveEducation #AIInEducation #CommunityEmpowerment

1

See the launch of a new online seminar series on using speculative fiction in research, policy, and activism. 📽️🌍 Join the team as they explore futures shaped by AI, automation, synthetic biology, surveillance capitalism, space exploration, nanotechnology, and integration into the global industrial economy. Subscribe at AppliedAfricanSF.com to stay informed about upcoming events and posts. Don’t miss out on these insightful sessions! #SpeculativeFiction #Research #Policy #Activism #AppliedAfricanSF

5

I first encountered the Theory of Jobs To Be Done from Arindam Mukherjee our Mentor at the NextLeap Product Manager Fellowship. Last night, I began plowing through the wonderful book, "The Prosperity Paradox: How Innovation Can Lift Nations Out of Poverty" co-authored by Clayton M. Christensen, Efosa Ojomo and Karen Dillon, from Harvard University, and then I met with the concept again. It was animating to read how this simple but marvellous theory has produced economic prosperity for nations, and how it underpins market-creating innovations. As a Product Manager with a heart for innovation and prosperity in Africa and emerging markets, I will be deepening my grasp of this theory and ensuring it becomes ingrained in my tool box. I have attached an excerpt from the book, one of the parts that talks about the theory and its role in innovation, but I have also outlined the key points below. Do have a read: 💡 Innovation and Consumption Economy - Innovation is often unreliable because it relies on past data about consumer behavior. - This data lacks insight into why people make choices and why they might avoid a purchase. ✅ Theory of Jobs to Be Done - People buy products to fulfill specific "jobs" in their lives. - Traditional marketing segments (demographics, psychographics) miss these underlying reasons. 🔨 Practical Example - Example: Buying the New York Times for specific purposes (e.g., avoiding a chatty neighbor, checking sports news). - Demographics don’t capture these specific "jobs." 📖 Importance of Understanding the Job - To ensure successful innovation, understand the specific "jobs" customers hire products to do. - Market-creating innovations solve unfulfilled "jobs" that had inadequate or no solutions. 📚 Case Studies study - Celtel: Solved the communication challenge in remote areas. - MicroEnsure: Provided insurance solutions where no adequate options existed. 💡 Nonconsumption Insight - People prefer nonconsumption over unsatisfactory solutions. - Understanding unfulfilled "jobs" reveals market potential for innovation. #Innovation #Prosperity #Markets #ProductManagement #ProductManager

20

Founder of the Distributed AI Research Institute (DAIR) and a leading voice in ethical AI, Dr. Timnit Gebru is reshaping how we think about accountability in artificial intelligence. Her advocacy ensures fairness, transparency, and equity are at the forefront of AI development. I AM THE ALGORITHM ensuring fairness and humanity remain at the core of AI development. Let’s celebrate Dr. Gebru for her trailblazing work in holding technology accountable to humanity. #IAmTheAlgorithm #BlackExcellenceInAI #EthicalAI #TechEquity #ResponsibleTech

3

As part of IDEAL Institute's Workshop on Navigating the Ethics of Human-AI Collaboration, Lionel Robert joins a panel on AI Agents in Business. Wednesday, 11/20 at 10:30am. More on the workshop: The rapid integration of AI systems into various domains of human life has sparked both enthusiasm and concern. While some envision unprecedented productivity gains, economic growth, and improved healthcare access, others worry about unfair discrimination, workforce displacement, and threats to human autonomy. Treating AI as a tool for augmenting human capabilities rather than replacing them, could reap the benefits of AI while preserving human agency and values. AI Agents (i.e., AI systems that can pursue complex goals with limited direct supervision) hold great promises for the augmentation of human capabilities. However, the emergence of more sophisticated and capable AI agents, which may play increasingly large roles in human lives, raises new questions about the nature of this augmentation and the evolving relationship between humans and AI. And register: https://lnkd.in/gfzG-hQi

14

1 Comment

Is it already time to decolonize AI? (Hint: of course!). It's eye-opening to see how unchecked biases can become embedded in our systems, especially in technology that shapes our perceptions and decisions. This isn’t just about faulty algorithms; it's about the responsibility we all share to create systems that genuinely reflect and respect the diverse world we live in. Decolonizing AI isn’t just a buzzword; it’s about rethinking whose voices are heard and valued! If we’re serious about creating a future rooted in equity and growth, we need to start asking tougher questions and seeking real change. It’s work worth doing.

2

3 Comments

More relevant than ever, Power On! explores Computer Science and its dangerous role in exacerbating inequalities and political polarization. Check out our blog featuring this MIT Press graphic novel and some thoughts from its creators, Jane Margolis and Jean Ryoo. ”We decided to focus on real-world problems that youth told us are pressing today, and we shaped a story around taking a critical stance about computing in our tech-saturated world. The main characters learn about racist AI; they find information about current tech workforce demographics and the lack of access to CS education in their school district. They also explore how CS is used across all fields and why everyone deserves to learn CS.” More in CLA Blog - Power On!: Why CS Education Can No Longer Pretend to Be “Neutral”: https://lnkd.in/gTd4JYrW #ComputerScience #EducationEquity #RacialEquity

6

Can a human masquerade as an AI? Here’s a fun video for your Friday, exploring a “reverse Turing test” scenario: Aristotle, Mozart, Leonardo da Vinci, Cleopatra and Genghis Khan are travelling together on an “AI only” ticket. The conductor knows one of the travellers is human. But which one? https://lnkd.in/e-axP6_A LLM powered NPCs are starting to be seen more often in games, and there are plenty of emerging healthtech applications that do the same (see Natal Mind for example). The ability to instruct models to adopt roles and debate themselves, or as agents, each other, also opens up fascinating opportunities to bring in diverse perspectives to clinical dilemmas. Of course, let’s hope we don’t have to pretend to be AI just to get let into the clinic in future, right? Have a great weekend! #AI #LLM #HealthTech #Agents #TuringTest

9

3 Comments

Harvard has released a great (and free) guide for any early stage founder. Highly recommend downloading a copy. #founder #founders #startup #startuplife #entrepreneur #entrepreneurship #businessowner #ceo #hustle #success #innovation 

1

*AI for good tldr*: Both training dataset size and model size influence intersectional bias (e.g. who gets labeled a "human" and/or "criminal" and/or "gorilla"). And yikes--decreasing one form of bias can lead to increases in other biases! Thanks to the amazing Dr. Abeba Birhane, the research team, and Mozilla for the careful and nuanced take on dataset scaling, models, and racial bias. (First, what is dataset scaling? A big idea in #ai is that more data is always better and results in great prediction accuracy. This can be achieved by either, one, getting more new data into your dataset or, two, by transforming your existing data through augmentation, synthetic data, etc. Pssst: this article is about the first use of dataset scaling). 👩🏿‍🎓👩🏻‍🎓👩🏽‍🎓 What can we learn from this research? 👩🏿‍🎓👩🏻‍🎓👩🏽‍🎓 👉🏻 Scaling dataset size increases racial bias in larger visio-linguistic models, particularly against Black and Latino men. For example, when the dataset was scaled from 400M to 2B samples, the probability of a larger ViT-L model predicting an image of a Black man as a criminal increased by 65%. 👉🏻 Smaller models exhibit reduced bias with larger datasets, while larger models show increased bias. For example, for the smaller ViT-B models, the probability of predicting an image of a Latino man as a criminal decreased by 47% when the dataset size increased from 400M to 2B samples. 👉🏻 Biases in AI models reflect deeper historical and societal issues, emphasizing the need for careful dataset curation and auditing. For example, scary sentences in here with a low bar: "For all racial groups, the probability of an image of a human from the CFD being predicted as human being was higher in smaller models (ViT-B-16 and ViT-B-32). On the other hand, the probability of an image of a human from the CFD being predicted as human being decreased for Latino women, Black women, White women, and White men in the larger models (ViT-L-14)." (8) 💡 Interesting insight: 💡 Despite reduced non-human offensive labels (like "gorilla") with larger datasets, models still misclassify Black individuals with offensive human labels (like "criminal"). This indicates that while overt dehumanization decreases, harmful stereotypes persist in new forms. The shift underscores the ongoing challenges in addressing bias in AI training data. 🚜 How do we get to work? 🚜 ☑ Are your data scientists tracking these sociotechnical ways of auditing AI systems for social outcomes? Or are they heads down on the technical details? Help them do both. ☑ Audit your own models! This is especially important for the social impact sector where our goals for equality and equity are foregrounded instead of profit. ☑ Ask your team to demonstrate, pre-Go Live, model performance by intersectional social groups and audit results assessing both overt and covert biases in your context/use case.

11

1 Comment

Earlier this month, #Hackweek brought together 18 talented program fellows, top faculty members and four industry sponsors to dive into intensive problem-solving sessions and collaborative projects that leverage the power of #AI. Hackweek is a cornerstone of ISEA: A Data Science Training Program to Advance Educational Research and Practice, supported by the Institute of Education Sciences (IES), UW College of Education and the UW eScience Institute. The ISEA program, led by Professor and Principal Investigator Dr. Min Sun, is designed to bridge the gap between AI and data science skills in K-12 education settings. Learn more about ISEA and Hackweek: https://lnkd.in/gdiJtKNF Hackweek participants engaged in a forward-thinking panel discussion with IES Commissioner Elizabeth Albro, PhD, various WA school district representatives and other educational organizations led by Dr. Sun, titled "Envision Generative AI Use Cases in K-12 Education and Capacity Building Strategies." This conversation highlighted the transformative potential of AI in schools, emphasizing its role in enhancing teaching efficiency, expanding access to quality education, especially in rural areas, and fostering deeper, human-centered interactions between teachers and students. Throughout the week, the teams used cutting-edge data analytics, qualitative and quantitative research methods, and innovative study design to advance real-world projects. The projects addressed during Hackweek included developing predictive models to improve educational outcomes and designing tools to facilitate real-time assessment of educational data. This was an excellent opportunity for everyone to apply insightful research techniques and creative approaches to tackle pressing educational issues. The outcomes of these intensive projects will be shared with a broader audience during a Fall Virtual Conference. #UWInnovates

13

Intersection of any general purpose technology with society and economy has been an important mark of development or progress or change, and many more. Technology that receives wide spread acceptance and usage is often the one that reduces cost, reduces time and increases convenience. Considering numbers, which also is an idea, move from abacus to Slide rules to logarithmic Tables, electronic calculators and later the spread sheets, data bases and relational databases changed the paradigm of number crunching. Words, again as ideas, with probably more wider and diverse logical and contextual relationships, can now be generated with the current version of Gen AI, based on matrix transformation in Large Language Models based on the deep learning of Neural network technologies. A convenient " Next word predictor" or now the "next sentence" or "next paragraph" predictor , depending upon the context of the "prompts", though fascinating need not be taken as miracle. Its simply Pure maths!! No magic to be over awed with. The ludite version of destruction of steam driven mechanized machine in late 1700s, is well known. The response, especially from the humanities side, often is of absolute distrust and rejection of most technology. The works from political science or sociology study in inequalities tend to paint as an unknown and unpredictable demon and thus peddle the version of scaremongering. Technology remains a tool for the motives and objectives of humans. But offering a agency and motives to technology, even the autonomous AI kind, seems misplaced. The dynamics of stock market speculative investments, the meteoric rise of related stocks, start up economics, has also been confounding factor influencing the status of AI and society. We have seen huge speculative rises , sometimes by tens of thousand of percents of some stocks, and the Impact on forming public opinion cannot be neglected. The entertainment industry, with probably knowledge as a new entry in its arsenal, has used AI for maximum extraction of public interest, and thus profits, especially targeting on intense emotions , the unpredictable aspects this new technologies has. Pandering to emotions like Fear anxiety have always been dependable profit making strategy of "media fictions" as well as "science fiction" and most editors seem to follow the line. Thus media hypes , apart from reporting news, have a good and simple business logic. Gen AI is a simple technology, as simple as a "word creator". The greatness imposed upon it , may not remain so in matter of months, as we will take its functionalities as normal, or as we enhance our expectations, and move up the benchmark. Hopefully it should live upto the expectations, that of dramatically enhancing the productivity, which again has never been unprecedented, but as yet still to be clearly demonstrated. #AI #GenAI #LLM #technology #ICT United Nations UNITAC Hamburg

7

6 Comments