OpenAI is funding academic research into algorithms that can predict humans’ moral judgements. In a filing with the IRS, OpenAI Inc., OpenAI’s nonprofit org, disclosed that it awarded a grant to Duke University researchers for a project titled “Research AI Morality.” Contacted for comment, an OpenAI spokesperson pointed to a press release indicating the award is part of a larger, three-year, $1 million grant to Duke professors studying “making moral AI.”
Tuesday, December 10, 2024
Embracing change: research transformation in the age of AI - Times Higher Ed
As researchers confront important global issues, they face growing demands to provide timely and impactful solutions. However, challenges such as increasing expectations, limited funding and insufficient infrastructure can create obstacles. In response, many are embracing emerging technologies such as AI for support. Times Higher Education hosted a webinar on the topic – in partnership with Digital Science – to discuss the evolving role of research in academia and the transformative impact of new technologies on making research more open, inclusive and collaborative.
Monday, December 09, 2024
The blueprint for colleges and universities in the new world of work - Robert Donnelly & Niko Milberger, University Business
Many college graduates find themselves in jobs that don't align with their studies or degrees, a trend that is further exacerbated by the perpetual advances in artificial intelligence and technology. Many college graduates find themselves in jobs that don’t align with their studies or degrees, a trend that is further exacerbated by the perpetual advances in artificial intelligence and technology. The urgency for academia and higher education institutions to recognize the rapid changes in the job market and the need for immediate adaptation is pressing. The jobs of the past and the educational credentials required for those jobs are rapidly disappearing, necessitating a reevaluation of standards and methods as the new world of work evolves daily.
Opinion: 4 Keys to Unlocking the Power of GenAI in Higher Ed - Chad Bandy and Saravanan Subbarayan, GovTech
To turn the disruption of generative artificial intelligence into an opportunity, higher education leaders should focus on four important variables: policy, principles, strategy and collaboration. Each institution is unique and at different stages in their GenAI journey. Some could reasonably be described as thriving, readily embracing new technologies and possessing the means and resources to enact their agenda. Others are still striving to reach the necessary level of technological prowess to drive forward in 2025. Faculty at many institutions are being sought after to advise legislators and leaders in their communities about risks and opportunities with AI. Nevertheless, institutions are grappling with where the greatest impacts with the least number of risks are with AI in their operations. Regardless of where any institution finds itself in their GenAI journey, one thing is clear: Much work remains to fully harness the benefits of GenAI and navigate the intricate landscape it represents.
Sunday, December 08, 2024
OpenAI partners with Wharton for a new course focused on leveraging ChatGPT for teachers - Preston Fore & Jasmine Suarez, Fortune
Generative artificial intelligence is the elephant in the (class)room at schools nationwide. And while many students have largely caught on to its omnipresence, teachers are lagging behind. OpenAI, the parent of ChatGPT, is hoping to change the dynamic with a new partnership with one of the best business schools in the country. The goal is to empower educators to effectively bring generative AI into the classroom and maximize learning, says Leah Belsky, vice president and general manager of education at OpenAI. “Teachers and professors are an important node in both learning how they can transform pedagogy and transform the way people learn with AI,” says Belsky. The new class is co-taught by Lilach and Ethan Mollick of the University of Pennsylvania—who have dedicated their lives to AI education.
AI and the Job Market - Kim Isenberg, Forward Future AI
If we look at further predictions, such as that made by former Google CEO Eric Schmidt that either this year or next, the limits of context windows will be exceeded, and then look at the abilities that general agents already have today, then there is no way of knowing where we will be in a few years and what the impact on the world of work will be. Accordingly, as I mentioned at the beginning, I will repeat the analysis in 2025 to see how the data has changed and adapted. But until then, we can safely say that AI will have a significant impact on jobs worldwide and will destroy jobs. There is agreement on this. The only disagreement is about how strong this influence will be.
Saturday, December 07, 2024
Apple’s new AI-powered Siri? - Martin Crowley, AI Tool Report
Insiders at Apple HQ have leaked that, after 13 years, Apple is building a new Siri—codenamed LLM Siri—designed to rival OpenAI’s ChatGPT with ‘Advanced Voice’, and Google’s Gemini Live. The new voice assistant will be powered by Large Language Models, allowing it to engage in more natural-flowing, 2-way conversations with users, and understand and complete complex tasks, quicker and more effectively.
Q STAR 2.0 - new MIT breakthrough AI model IMPROVES ITSELF in REAL TIME (new Strawberry?) - Wes Roth, YouTube
The podcast is about a new AI model called Q* 2.0 that has the potential to be a significant breakthrough in artificial intelligence. The model is based on a technique called test time training (TTT), which allows it to adapt and improve its performance in real time as it is being tested. This is a significant departure from traditional AI models, which are typically trained on a large dataset and then evaluated on a separate test set. (summary developed with Gen AI assistance)
Friday, December 06, 2024
This New AI Model Is Genius - DESTROYS OpenAI o1 in REASONING - AI Revolution
DeepSeek's R1 AI model is making waves in the AI field with its advanced reasoning abilities, rivalling and even outperforming OpenAI's 01 model in certain areas. R1 utilizes Chain of Thought reasoning to break down complex tasks into smaller steps, excelling in accuracy and reliability, particularly in solving math problems and real-world scenarios. It also boasts transparency by revealing its reasoning process, unlike other AI models. Despite facing challenges with specific logic problems and potential for misuse (jailbreaking), R1 is a significant advancement, backed by substantial investment and a commitment to open-source development. This development is part of a broader trend in AI, moving towards more refined reasoning and personalized user experiences. (summary assisted by Gen AI)
AI chatbot can conduct research interviews on unprecedented scale - Juliette Rowsell, Times Higher Ed
Freely available tool performs strongly in trials against human interviewers and traditional online surveys. Two London School of Economics scholars have developed a chatbot powered by a large language model that, they say, can complete interviews with thousands of participants in a matter of hours. Rather than having a standard set of multiple-choice and open text questions, as has typically been the case with online surveys, the chatbot takes a conversational approach, collecting interviewees’ responses and using them to generate new questions within a broad set of parameters. Its creators, Friedrich Geiecke, an assistant professor of computational social science, and Xavier Jaravel, a professor of economics, say the tool employs best practice from academic literature – for example, encouraging participants to freely express their views, and then posing follow-up questions to ensure clarity.
Thursday, December 05, 2024
AI simulations of 1000 people accurately replicate their behaviour - Chris Stokel-Walker, New Scientist
Using GPT-4o, the model behind ChatGPT, researchers have replicated the personality and behaviour of more than 1000 people, in an effort to create an alternative to focus groups and polling. An experiment simulating more than 1000 real people using the artificial intelligence model behind ChatGPT has successfully replicated their unique thoughts and personalities with high accuracy, sparking concerns about the ethics of mimicking individuals in this way. Joon Sung Park at Stanford University in California and his colleagues wanted to use generative AI tools to model individuals as a way of forecasting the impact of policy changes. Historically, this has been attempted using more simplistic rule-based statistical models, with limited success.
Amazon launches Nova AI model family for generating text, images and videos - Carl Franzen, Venture Beat
The Amazon Nova suite introduces several models tailored to specific use cases, all supporting more than 200 languages:
• Amazon Nova Micro: A text-only model optimized for low-latency responses at minimal cost.
• Amazon Nova Lite: A multimodal model offering fast processing for text, images, and videos at a very low cost.
• Amazon Nova Pro: A multimodal model combining accuracy, speed, and cost-efficiency, designed for a wide range of tasks.
• Amazon Nova Premier: The most advanced multimodal model for complex reasoning tasks and for distilling custom models (launching in Q1 2025).
• Amazon Nova Canvas: An advanced image generation model for creative content development.
• Amazon Nova Reel: A state-of-the-art video generation model offering dynamic capabilities.
All models support fine-tuning and knowledge distillation, allowing customers to tailor AI tools to their proprietary data for improved accuracy and performance. These models excel in supporting Retrieval Augmented Generation (RAG), which grounds outputs in specific organizational data to enhance reliability.
Wednesday, December 04, 2024
Orchestrator agents: Integration, human interaction, and enterprise knowledge at the core - Emilia David, Venture Beat
There is no doubt AI agents will continue to be a fast-growing rend in enterprise AI. But as more companies look to deploy agents, they’re also looking for a way to help them make sense of the many actions these autonomous or semi-autonomous, AI guided bots will take, and avoid conflicts. Enter the orchestrator: these type of agents function as managers of other, more specialized agents, understanding each one’s role and activating each based on the next steps needed to finish a task. Most orchestrator agents, sometimes called meta agents, monitor if an agent succeeded or failed and choose the following agent to trigger to get the desired outcome.
Higher Education in 2025: AGI Agents to Displace People - Ray Schroeder, Inside Higher Ed
The new year may bring a host of virtual assistants and administrative staff to higher education. They will begin as assistants to humans, then over time they will evolve into autonomous AI staff members. Next-generation agents are in development and near release. As Bernard Marr reports in Forbes, “While previous iterations of AI focused on making predictions or generating content, we’re now witnessing the emergence of something far more sophisticated: AI agents that can independently perform complex tasks and make decisions. This third wave of AI, known as ‘agentic AI,’ represents a fundamental shift in how we think about and interact with artificial intelligence in the workplace.”
Tuesday, December 03, 2024
Google’s Gemini has a memory: Google’s Gemini can now remember what you tell it - Martin Crowley, AI Tool Report
Following OpenAI’s launch of ChatGPT’s memory feature in April, Google has revealed that its own chatbot, Gemini, can now “remember the things you care about” such as your life, work, aspirations, and personal preferences, and, as a result, delivers more tailored and relevant responses. For example, if you tell Gemini you’re a Vegan, and then ask it for restaurant recommendations, it will only display places that are suitable for Vegans. Only users on Google’s Google One AI Premium subscription plan have access to this memory feature, and it's currently only available in English. They can add their interests and preferences through the chatbot interface, or via the “saved info” page, where Google has provided some examples of what they might like Gemini to remember, like “Use simple language and avoid jargon” or “When planning a trip, always include the cost per day,” for example.
How Are Companies Really Using AI? A New Report Has Answers - Stefano Puntoni, Knowledge at Wharton
“Growing Up: Navigating Gen AI’s Early Years” is a survey of more than 800 senior business leaders in large organizations that reveals a seismic shift in their attitudes and applications of AI in just a short time. In 2023, the first year of the survey, only 37% reported using AI weekly. That number has risen to 72% in 2024. Negative perceptions, namely worry and skepticism, are softening as decision-makers explore how this evolving technology can help their firms become better. According to the survey, generative AI is being widely deployed across functions, even in departments such as marketing and human resources that were initially slower to adopt it. The highest use is in document and proposal writing and editing with 64%, followed closely by data analyses and analytics at 62%. Other high-use functions include customer service and support (58%), fraud detection and prevention (55%), and financial forecasting and planning (53%).
Monday, December 02, 2024
Is Algorithmic Management Too Controlling? - Lindsey Cameron, Knowledge at Wharton
In more and more workplaces, important decisions aren’t made by managers but by algorithms which have increasing levels of access to and control over workers. While algorithmic management can boost efficiency and flexibility (as well as enabling a new class of quasi-self-employed workers on platforms like Uber and Instacart), critics warn of heightened surveillance and reduced autonomy for workers. In a newly published paper, Wharton Prof. Lindsey Cameron examines how ride-hail drivers interact with the algorithmic management tools that make app-based work possible. In this interview, she shares insights from her research, along with tips for creating a more equitable future of work.
https://knowledge.wharton.upenn.edu/article/is-algorithmic-management-too-controlling/
Deep learning pipeline for accelerating virtual screening in drug discovery - Fatima Noor, et al; Nature Scientific Reports
In benchmarking, VirtuDockDL achieved 99% accuracy, an F1 score of 0.992, and an AUC of 0.99 on the HER2 dataset, surpassing DeepChem (89% accuracy) and AutoDock Vina (82% accuracy). These results underscore the tool’s capability to identify high-affinity inhibitors accurately across various targets, including the HER2 protein for cancer therapy, TEM-1 beta-lactamase for bacterial infections, and the CYP51 enzyme for fungal infections like Candidiasis. To sum up, VirtuDockDL combines user-friendly interface design with powerful computational capabilities to facilitate rapid, cost-effective drug discovery and development. The integration of AI in drug discovery could potentially transform the landscape of pharmaceutical research, providing faster responses to global health challenges. The VirtuDockDL is available at https://github.com/FatimaNoor74/VirtuDockDL .
Sunday, December 01, 2024
Google’s New AI Is Shockingly Good and Scary - AI Revolution
Google's new AI model, Gemini x114, is making waves in the AI field. It has achieved the top spot on the Chatbot Arena leaderboard, outperforming even OpenAI's GPT-4. This model excels in various tasks, including math, creative writing, and visual understanding. However, access to Gemini x114 is currently limited to developers and researchers through Google AI Studio. The model's success highlights the ongoing debate about how to measure AI progress, as its performance drops significantly when factors like response formatting are controlled. Beyond its technical capabilities, Gemini x114 has also sparked controversy due to instances of generating troubling and insensitive responses. (summary provided by Gen AI)
If AGI arrives during Trump’s next term, ‘none of the other stuff matters’ - Harry McCracken, Fast Company
I think it depends on the extent to which Donald Trump will listen to Elon Musk. On one hand, you have a lot of folks who are very anti-regulation trying to persuade Trump to repeal even Biden’s executive order, even though that was very weak sauce. And then on the other hand, you have Elon, who’s been pro AI regulation for over a decade and came out again for the California regulation, SB 10 47. This is all going to really come down to chemistry and then relative influence. In my opinion, this issue is the most important issue of all for the Trump administration, because I think AGI is likely to actually be built during the Trump administration. So during this administration, this is all going to get decided: whether we drive off that cliff or whether AI turns out to be the best thing that ever happened.
Subscribe to:
Posts (Atom)