Latest news with #LargeLanguageModels
Yahoo
2 days ago
- Business
- Yahoo
New York Times, Amazon Unveil AI Content Licensing Deal
In its first AI licensing deal, The New York Times Company and giant Amazon have announced a multi-year agreement that will bring Times editorial content to a variety of Amazon customer experiences in a move the partners said broadens the companies' existing relationship. The idea is to 'bring additional value to Amazon customers and bring Times journalism to wider audiences,' the companies said. More from Deadline Netflix Co-Founder Reed Hastings Joins Board Of AI Firm Anthropic Amazon's Head Of Unscripted Series Jenny Falkoff Joins Netflix Darren Aronofsky's AI-Focused Studio Primordial Soup Sets Strategic Partnership With Google DeepMind Under the new deal, Amazon is licensing editorial content from The New York Times, NYT Cooking, and The Athletic 'for AI-related uses.' This will include real-time display of summaries and short excerpts of Times content within Amazon products and services, such as Alexa, and training Amazon's proprietary foundation models. The collaboration will make The New York Times's original content more accessible to customers across Amazon products and services, including direct links to Times products, 'and underscores the companies' shared commitment to serving customers with global news and perspectives within Amazon's AI products.' As AI firms suck up vast quantities of data to train their so-called Large Language Models publishers have taken different tacks, some inking deals some seeking the courts. The NYT is suing giant OpenAI and its major investor Microsoft for copyright violation in its use of content. A judge ruled in March that the suit can proceed. Copyright is the issue and 'fair use,' a legal doctrine that allows use of copyrighted material in certain ways and in certain cases. AI companies have often appeared game to compensate publishers on their own terms but are seeking loser restrictions on copyright rules in order to grown and, they said in recent testimony, compete globally. Copyright owners including the Hollywood creative community have pushed back on that, insisting that the laws be upheld. It's not clear if or how turmoil at the U.S. Copyright Office will impact this. A federal judge yesterday declined to issue an order that would immediately prevent the Trump administration from firing the register of copyrights and head of the office, Shira Perlmutter Best of Deadline Everything We Know About Netflix's 'The Thursday Murder Club' So Far 2025 TV Series Renewals: Photo Gallery 2025-26 Awards Season Calendar: Dates For Tonys, Emmys, Oscars & More Error in retrieving data Sign in to access your portfolio Error in retrieving data Error in retrieving data Error in retrieving data Error in retrieving data
Forbes
3 days ago
- Health
- Forbes
Protecting Your Mind Amid AI's Persuasive Power Play
In the marketplace of ideas, from political campaigns to product marketing, persuasion has long been a human art form. We rely on logic, emotion, charisma, and trust to influence and be influenced. But a new power player is rapidly entering the fray: Artificial Intelligence. Sophisticated AI, particularly Large Language Models are no longer just information processors; they are becoming skilled digital persuaders, capable of shaping opinions and nudging behaviors in ways we are only beginning to understand. The question is no longer if AI can be persuasive, but how persuasive it can be, and what that means for our future. The foundations of human persuasion are well-documented, perhaps most famously by Dr. Robert Cialdini, who outlined principles like reciprocity, scarcity, authority, commitment and consistency, liking, and social proof. These psychological levers have been the bedrock of influence strategies for decades. Humans excel at deploying these intuitively, building rapport, reading nuanced social cues, and leveraging genuine emotional connections to build deep, lasting trust. However, the digital age has ushered in AI systems with a distinct set of advantages. These algorithms can process and analyze vast datasets on human behavior, preferences, and communication styles, allowing for an unprecedented level of personalized messaging at scale. Imagine an AI that can tailor its arguments and tone in real-time, A/B testing thousands of variations of a message to find the most effective one for a specific individual or demographic – a feat impossible for a human. Recent studies underscore this emerging reality. Research has shown that AI-generated messages can be as, or in some cases even more, persuasive than those crafted by humans. Making them significantly more effective in changing minds on divisive topics in online debates. Simply making models bigger doesn't inherently make a single message dramatically more influential, but the overall trend indicates a powerful new persuasive force. One compelling example of this specialized persuasive technology comes from academia. The paper AI-Persuade: A Conversational AI for Persuasion Towards Pro-Environmental Behaviors details a system designed specifically to influence users to adopt more environmentally friendly habits. This AI doesn't just present facts; it engages in interactive conversations, employing a diverse toolkit of persuasion strategies — such as goal setting, positive framing, and social commitment — to foster long-term attitudinal and behavioral shifts. The researchers' user studies validated its potential to effectively guide individuals towards targeted outcomes. This points to a future where AI could be a significant force in public service campaigns, health interventions, and educational initiatives. AI's persuasive power isn't just about brute-force data processing. It taps into several psychological mechanisms: Despite AI's growing capabilities, human interaction retains unique strengths in persuasion. Genuine empathy, the ability to understand and share the feelings of another, is profoundly difficult for AI to replicate authentically. Building deep, long-term trust, the kind that underpins significant life changes or high-stakes decisions, often relies on shared experiences, vulnerability, and the nuanced dance of human relationships. Humans can adapt to entirely novel situations with a flexibility and intuition that current AI lacks, drawing on a lifetime of complex social learning. It matters to remember that AI is a tool to an end. The latter must be decided up by human users, based on ethics and moral values. The same tools that can encourage positive behaviors may be weaponized for manipulation, spreading misinformation, or unduly influencing vulnerable populations. The potential for AI-generated propaganda or highly personalized, deceptive marketing campaigns is a serious concern that demands ethical guidelines, transparency in AI deployment, and a focus on media literacy. AI's impact on decision-making and overreliance on our artificial assistants can diminish critical thinking, making us susceptible to manipulation if we're not vigilant. Ultimately, the good and bad of AI depends on the human mindset. The future likely involves a hybrid landscape where AI and human persuasion coexist and even collaborate. AI might handle initial engagement, provide personalized information, or manage large-scale outreach, while humans step in for more complex, empathetic, and high-trust interactions. As AI's persuasive abilities become more integrated into our lives, we need a framework to navigate this new terrain responsibly and effectively. Consider the A-Frame: The rise of the digital deluge is upon us. By understanding its power, recognizing its mechanisms, and committing to a framework of mindful engagement, we can harness the benefits of persuasive AI while safeguarding our autonomy and critical judgment in an increasingly AI-influenced world.
Yahoo
4 days ago
- Business
- Yahoo
Deutsche Bank Boosts Digital Transformation With Collaborations
Deutsche Bank DB has embarked on a digital transformation drive, focusing on cloud migration, artificial intelligence (AI), and automation to enhance operational efficiency and client services. In sync with this, this week, DB has collaborated with International Business Machines IBM and finaXai, a Singapore-based AI company. DB announced that it reinforced its strategic partnership with IBM through a new license agreement, providing the former with greater access to International Business Machines' advanced software solutions. By leveraging IBM's innovative technologies, Deutsche Bank aims to streamline workflows, reduce operational costs and improve efficiency across all areas of its business. The transition from legacy systems to IBM's advanced cloud and AI solutions will allow DB to develop a more agile, scalable and secure technology stack. IBM's technologies, including the watsonx AI portfolio, will enable Deutsche Bank to deliver more personalized and seamless services to clients, from smarter customer support to tailored financial solutions. The integration of IBM's AI capabilities is expected to not only improve operational efficiency but also enhance fraud detection, risk management and regulatory compliance, leading to substantial cost savings and revenues. Deutsche Bank has announced a strategic partnership with finaXai, co-founded by researchers from Nanyang Technological University and the National University of Singapore. This partnership is expected to significantly expand Project DAMA 2 — Deutsche Bank's public-permissioned multi-chain asset servicing pilot. Notably, Project DAMA 2 intends to transform the management and servicing of tokenized funds and their investors by leveraging digital and AI technologies. Deutsche Bank and finaXai are set to explore the integration of technologies, including Machine Learning and Large Language Models, into asset servicing workflows. The focus will be on leveraging explainable AI to enable asset managers to execute fund lifecycle activities with enhanced speed, precision and transparency. Anand Rengarajan, Deutsche Bank's head of securities services APAC & MEA and global head of sales, stated, 'We are delighted to work with a diverse mix of experts from the academic and innovation worlds to contribute to the future of tokenised assets.' 'Through this collaboration, we unify leading research-backed solutions with industry applications to reduce complexity and boost AI's explainability and integration with DLT. With this work, we seek to better understand and anticipate our asset manager clients' needs as they explore the future of tokenisation,' he added. These strategic collaborations emphasize Deutsche Bank's commitment to adopting innovative technologies to reshape the future of financial services. By broadening its partnerships with IBM and finaXai, DB is not only transforming its technology stack and operational workflows but also setting new standards in digital asset servicing. Through these initiatives, the bank is positioned to enhance efficiency, security and client experience while navigating the evolving landscape of tokenized assets and AI-driven financial services. Over the past six months, DB shares have surged 64.7% on the NYSE compared with the industry's rise of 25%. Image Source: Zacks Investment Research Currently, Deutsche Bank carries a Zacks Rank #3 (Hold). You can see the complete list of today's Zacks #1 Rank (Strong Buy) stocks here. This month, Citigroup Inc. C unveiled Citi AI, a range of AI tools aimed at enhancing internal processes for Hong Kong employees. Citi AI aims to maximize efficiency in operations by offering support in information retrieval, document summarization and writing electronic communications for employees. Citigroup AI is currently accessible to approximately 150,000 employees across 11 countries, including the United States, India and Singapore. The company plans to expand availability to more markets this year. In November 2024, PNC Financial Services Group, Inc. PNC announced its partnership with GTreasury, a global leader in Digital Treasury Solutions. This collaboration integrated PNC's embedded banking solution, PINACLE Connect, with GTreasury's treasury and risk management platform. PNC introduced PINACLE Connect in 2021 to remove complexities and address client feedback, particularly regarding technology integration and fraud prevention. Want the latest recommendations from Zacks Investment Research? Today, you can download 7 Best Stocks for the Next 30 Days. Click to get this free report Citigroup Inc. (C) : Free Stock Analysis Report The PNC Financial Services Group, Inc (PNC) : Free Stock Analysis Report International Business Machines Corporation (IBM) : Free Stock Analysis Report Deutsche Bank Aktiengesellschaft (DB) : Free Stock Analysis Report This article originally published on Zacks Investment Research ( Zacks Investment Research Sign in to access your portfolio
Finextra
5 days ago
- Business
- Finextra
Deutsche Bank collaborates with FinaXai to adopt AI for tokenised funds servicing
Deutsche Bank is collaborating with finaXai, a Singapore-based AI company co-founded by researchers from Nanyang Technological University, Singapore (NTU Singapore) and the National University of Singapore (NUS). 0 The collaboration will assess a significant extension of Project DAMA 2, the public-permissioned multi-chain asset servicing pilot focused on transforming how asset managers can more efficiently manage and service tokenised funds and their investors. FinaXai is also part of the Fincubator programme at the Asian Institute of Digital Finance (AIDF), a university-level institute in NUS jointly established by the Monetary Authority of Singapore, the National Research Foundation and NUS, to support the development of fintech startups focused on deep technology and digital finance innovation. The team will explore integrating Machine Learning and Large Language Models (LLMs) into asset servicing workflows, leveraging explainable AI to help asset managers plan and execute fund lifecycle activities with greater speed, transparency, and precision. Anand Rengarajan, Deutsche Bank's Head of Securities Services APAC & MEA and Global Head of Sales said: 'We are delighted to work with a diverse mix of experts from the academic and innovation worlds to contribute to the future of tokenised assets. Through this collaboration, we unify leading research-backed solutions with industry applications to reduce complexity and boost AI's explainability and integration with DLT. With this work, we seek to better understand and anticipate our asset manager clients' needs as they explore the future of tokenisation.' Dr. Erik Cambria, Co-Founder of finaXai and Professor, College of Computing and Data Science, NTU said: "finaXai is excited to partner with Deutsche Bank in applying cutting-edge explainable AI techniques to the domains of finance and asset tokenisation. This collaboration bridges academic research with real-world applications and explores how explainable AI can converge with tokenisation to streamline processes and enhance the accessibility, adoption, and management of digital assets. Ultimately, this initiative lays the groundwork for synergies between trustworthy AI and asset managers, enabling faster, more precise planning and execution of digital asset and fund lifecycle activities." Dr. Gianmarco Mengaldo, Co-Founder of finaXai and Assistant Professor at the Department of Mechanical Engineering, College of Design and Engineering, NUS, said: 'finaXai operates at the exciting intersection of AI applications in finance and comprises researchers from various institutions who share a common interest in exploring the frontiers of AI in this field. This collaboration demonstrates the significant impact that can be achieved when scientific interests align. The team based in NUS focuses on bridging human understanding with AI methods through explainable AI, and integrating existing knowledge into AI-driven solutions when beneficial. It will be exciting to see how our work, complemented with the work by NTU counterparts, can aid in dealing with complex real-world tasks in the financial sector.' Representatives involved in the project include: Deutsche Bank: Boon Hiong Chan, Asia Pacific Head of Securities & Technology Advocacy and Industry Applied Innovation Lead. Jie Yi (Jaelynn) Lee, Digital Product Owner. finaXai: Dr. Erik Cambria, Co-Founder of finaXai and Professor, College of Computing and Data Science, NTU. Dr. Gianmarco Mengaldo, Co-Founder of finaXai and Assistant Professor at the Department of Mechanical Engineering, College of Design and Engineering, NUS. Dr. Mao Rui, Technology Lead and Research Scientist, College of Computing and Data Science, NTU. Keane Ong, Technology Scientist and PhD candidate in Digital Financial Technology, specialising in Machine Learning and Natural Language Processing at NUS. Federico Cristina, CFA, Business Lead, finaXai.
Time Business News
6 days ago
- Business
- Time Business News
Nihar Malali's Vision for AI-Driven Transformation in Life Insurance: Bridging Innovation and Ethical Risk Assessment
As the insurance industry navigates a paradigm shift propelled by artificial intelligence (AI) and machine learning (ML), few professionals have played as crucial a role in shaping its direction as Nihar Malali. Currently serving as a Principal Solutions Architect, Malali brings more than two decades of multifaceted experience in cloud computing, AI-powered platforms, and enterprise integration. His leadership is helping redefine the architecture of digital transformation across financial services, particularly in life insurance claims adjudication and underwriting. At the intersection of two of his recent research endeavors lies a powerful and timely synthesis: how advanced models such as Large Language Models (LLMs), Retrieval-Augmented Generation (RAG), and neural networks can radically improve the accuracy, efficiency, and fairness of life insurance claims processing and underwriting. Rethinking Claims Adjudication with LLMs and RAG Traditionally, life insurance claims adjudication has depended heavily on manual review processes, rule-based systems, and actuarial tables. These methods, while foundational, are increasingly challenged by the rising complexity and volume of claims data. In his recent paper on AI-powered claims adjudication, Nihar Malali introduces a new paradigm in which LLMs and RAG architectures work in unison to automate, accelerate, and enhance claims decision-making. These architectures allow insurers to go beyond surface-level data entry and into intelligent, real-time risk assessment. LLMs are capable of analyzing unstructured documents such as hospital reports, policy agreements, and witness statements, extracting critical information for faster and more accurate claims validation. When augmented with RAG systems, these models gain the ability to access and retrieve contextually relevant external data in real time, enhancing their decision-making capabilities and reducing hallucinations that often plague generative models. For Malali, this is not merely a technological upgrade—it's a strategic imperative. 'By integrating LLMs and RAG, we're equipping insurers with tools to cut down on turnaround time, improve fraud detection, and ensure regulatory compliance, all while enhancing the customer experience,' Malali noted in a recent LinkedIn thought piece. In practical terms, RAG architectures work by enhancing the LLM's prompt with retrieved, relevant documents. For example, when adjudicating a claim involving a rare medical condition, the system can pull data from approved medical literature or the insurer's knowledge base before generating a final recommendation. This improves transparency and fosters confidence among stakeholders. The Role of Predictive Modeling and RPA Malali's framework doesn't stop at language models. He also highlights the synergistic use of Robotic Process Automation (RPA) and predictive modeling. RPA is ideal for managing repetitive administrative tasks—like eligibility verification and data validation—freeing up human experts for complex decisions. Meanwhile, predictive models trained on past claims data can forecast likely outcomes or flag anomalies that might indicate fraud. This confluence of AI technologies represents a milestone in operational transformation, enabling insurers to balance accuracy, scalability, and customer trust. For Malali, who has worked extensively with Microsoft Azure, AWS, and GCP in deploying secure and scalable cloud infrastructures, this kind of layered AI architecture is where the future of insurance resides. Elevating Risk Assessment with Neural Networks in Underwriting Complementing his work in claims adjudication is Nihar Malali's second research paper on AI in life insurance underwriting. In it, Malali examines how machine learning models—particularly neural networks—can outperform traditional actuarial approaches in assessing client risk. Using a real-world dataset of over 15,000 anonymized life insurance applications, the study evaluated the efficacy of Random Forest (RF), Stochastic Gradient Descent (SGD), and neural network models across key performance indicators like precision, recall, and F1-score. The neural network emerged as the most robust, achieving 98% accuracy and 99% F1-score. The results demonstrated the model's reliability in identifying high-risk individuals without false negatives—an essential quality in ensuring fairness in coverage decisions. What sets Malali's work apart is not just its technical sophistication but its ethical awareness. He places equal emphasis on fairness, data governance, and model transparency—factors often overlooked in performance-centric AI initiatives. Through techniques like SMOTE (Synthetic Minority Over-Sampling Technique) for class balancing and rigorous imputation for missing values, Malali's preprocessing pipeline ensures that minority populations and edge cases are not misrepresented or excluded from risk assessments. From Accuracy to Accountability: Ethical AI in Practice One of the standout contributions from Malali's underwriting research is his critical engagement with the ethical implications of AI. As algorithmic decision-making becomes more entrenched in the underwriting lifecycle, concerns around bias, explainability, and data privacy have taken center stage. In his proposed framework, Malali suggests embedding explainable AI (XAI) methods into underwriting systems to help human analysts understand how and why decisions are made. This is particularly important in edge cases—such as applicants with non-traditional work histories or undocumented health conditions—where black-box models may default to rejection without clear justification. 'Ethical AI is not just a compliance checkbox. It's about maintaining public trust and ensuring that innovations serve everyone equally,' Malali has remarked in internal presentations to industry stakeholders. This perspective aligns with his leadership style, which combines technical vision with human-centric design. As a mentor to emerging engineers and a steward of governance processes like architecture review boards and CI/CD pipelines, Malali ensures that AI deployments are not only powerful but also responsible. Unified Impact: Bridging Automation and Empathy Bringing both strands of his research together, Nihar Malali envisions a life insurance industry that is both technologically advanced and deeply humane. On the one hand, AI-powered claims adjudication platforms reduce delays and administrative overhead. On the other, ethically designed underwriting systems ensure that individuals are evaluated fairly and comprehensively. In a world where policyholders increasingly expect personalized, on-demand service, these advancements are more than backend optimizations—they redefine the relationship between insurers and their customers. From real-time quote generation to instant claim settlement, Malali's work points to an industry that is faster, fairer, and more transparent. The Road Ahead: A Call for Scalable and Responsible AI Looking forward, Malali emphasizes the need for scalable AI architectures that can adapt to new regulatory, ethical, and technological landscapes. He advocates for the integration of wearable technology data into underwriting models and the exploration of blockchain for audit trails and data integrity. Moreover, he champions continuous learning systems that evolve alongside changes in medical science, demographic trends, and behavioral data. This dynamic adaptability, coupled with strong ethical underpinnings, is the cornerstone of his vision for the future. As the insurance sector continues to evolve under the weight of digital transformation, leaders like Nihar Malali are ensuring that the shift is both strategic and principled. Through his groundbreaking research and real-world implementations, Malali exemplifies what it means to engineer not just smarter systems—but a more just and inclusive future for insurance. To explore further details about Nihar Malali's work, research publications, and continued contributions to the AI domain, please visit his professional LinkedIn profile page and ResearchGate profile. TIME BUSINESS NEWS
