Welcome to this weeks edition of Neurians AI Digest, our curated newsletter on all things Artificial Intelligence (AI).

Our AI digests will cover insights & updates on how AI is impacting industries such as Healthcare, Education, Finance, Retail, Travel & the Software industry itself.

Here’s a recap of this week in AI:

Healthcare:

Google Cloud has announced the general availability of its Vertex AI platform and new features for its Healthcare Data Engine (HDE), aiming to drive innovation at the intersection of AI and healthcare. Vertex AI, now enhanced, enables healthcare professionals to better query health records, derive insights, and perform advanced analytics. It integrates Google's healthcare-specific AI models, such as Gemini and MedLM, the latter having passed U.S. Medical Licensing Exam-style questions, marking a milestone for healthcare generative AI.

The HDE tool, now globally available, supports better interoperability of healthcare data, offering providers a unified view of patient information. A study by Google Cloud and The Harris Poll highlights the burden of administrative tasks on healthcare workers, noting that AI could alleviate these challenges, helping providers focus more on patient care. (Source: Forbes)

Google DeepMind has introduced a new AI model called Tx-LLM (Therapeutic Large Language Model) designed to enhance pharmaceutical research by predicting key properties of potential drugs. This model represents a shift toward specialized AI tools tailored to specific industries, which can be more effective in addressing complex challenges compared to general-purpose AI.

Tx-LLM employs fine-tuning, refining a pre-trained model on targeted datasets to improve performance in drug discovery tasks, such as molecule identification and forecasting clinical trial outcomes. By leveraging extensive drug-related data, Tx-LLM could significantly reduce the time and costs associated with bringing new drugs to market.

The implications of this approach extend beyond pharmaceuticals. Industries such as manufacturing and automotive are also exploring specialized AI to enhance efficiency and innovation. For instance, AI can optimize production lines, predict equipment failures, and accelerate vehicle design and testing.

High-stakes sectors like pharmaceuticals, finance, and transportation stand to gain the most from custom AI models. In pharmaceuticals, AI can streamline the path from discovery to market while navigating regulatory hurdles. In finance, it can provide competitive advantages by offering tailored algorithms for specific markets. In transportation, AI can optimize routes and enhance supply chain management.

Specialized AI models offer adaptability, allowing them to adjust to the unique needs of different industries while maintaining benefits like speed and accuracy. This adaptability facilitates "algorithmic knowledge transfer," where insights from one sector can improve processes in another, promoting cross-industry innovation. Overall, tailored AI solutions can significantly reduce time to market and drive innovation across various sectors. (Source: Pymts)

At Zoomtopia 2024, Zoom has introduced "Zoom Workplace for Healthcare," which includes the AI Companion 2.0 to assist healthcare personnel in increasing productivity. A custom AI Companion add-on is also available, allowing for personalization with healthcare-specific dictionaries and integration with third-party data sources like electronic health records (EHRs). Additionally, the "Zoom Workplace for Clinicians" offering focuses on streamlining clinical workflows by automating tasks. A key feature is the use of healthcare-specialized AI for generating clinical notes, reducing the burden of documentation and enhancing doctor-patient interactions. (Source: GlobeNews).

Rapid advances in ChatGPT that began outperforming medical students and professionals on medical exams, prompted Harvard Medical School (HMS) to incorporate AI into its teaching. This includes a one-month AI course for incoming students in the Health Sciences and Technology (HST) track, emphasizing the importance of AI, data, and machine learning in future healthcare. HMS also introduced an AI-focused PhD track due to growing demand. Additionally, the school launched Dean’s Innovation Awards, offering grants to projects that integrate AI into education, research, and administration. Hospitals affiliated with HMS, like Brigham and Women’s, are testing AI tools to reduce doctors' administrative burden, enhancing patient care. (Source: Harvard)

Education: 

Is AI use acceptable for students? A Hingham High school senior was accused of cheating on a team project set to compete in a worldwide youth history contest.

The parents of the High School senior are suing school officials, claiming their son's civil rights were violated after he was unfairly punished for using AI to research and outline the social studies project. The student’s lawyer argues that the use of AI was not prohibited by school policy or assignment guidelines, but the student was penalized with a low grade, harming his GPA and initially blocking his admission to the National Honor Society. The lawsuit seeks to have the student's academic record corrected by removing the infractions and changing his grade.

The case raises broader questions about how schools handle the use of AI in assignments, especially as AI becomes more prevalent. The school contends the punishment was meant as a "teachable moment," but the student received a "D" on a second project after being required to redo it individually. (Source: WBur).

When leveraged with the right governance in place however, AI can be extremely advantageous to educators.

Seventh grade math teacher Benjamin Hamill used an AI-powered tool from Kiddom to automate feedback on student exit tickets, significantly improving his workflow. Before using AI, Hamill struggled to provide timely feedback to his 120 students daily. The AI tool allowed him to offer personalized responses linked to educational standards, guiding students on how to improve. He noted that the feedback feature was highly effective, while automated grading was less reliable, with about 15-20% errors. Kiddom has since rolled out three AI tools—automated feedback, grading, and a practice problem generator—to help teachers save time. (Source: GovTech)

Zoom's new "Zoom Workplace for Education" suite includes the AI Companion feature to enhance educators' workflows with tools for lesson planning and real-time material review. Students can benefit from free AI-powered note-taking, which allows them to capture and summarize lectures on their phones, making it an appealing option for wide student adoption.

However, Zoom faces a challenge in balancing free student tools with revenue-generating features for educators, which focus more on administrative tasks. To succeed, Zoom may need to pivot towards broader institutional services, especially as it faces competition from established platforms like Blackboard, Canvas, and AI-driven tools like Course Hero. (Source: NoJitter)

Finance

Can BigBanks benefit from AI? What are the implications for workers? JPMorgan Chase CEO Jamie Dimon acknowledges that while AI has the potential to significantly boost productivity and reduce costs, it will also lead to job displacement, and banks need to take action to support affected workers. Faisal Husain, CEO of Synechron, which advises many of the world's largest banks on AI, agrees that AI will be transformative but stresses that we are still in the early stages of its development. He notes that the full impact on jobs will be gradual, and it will take time for banks to realize a return on AI investments.

Husain advises financial institutions to focus on small, productivity-enhancing applications of AI rather than searching for a "killer app." These incremental improvements can help integrate AI into enterprise operations more effectively. He emphasizes that AI is evolving quickly, but companies should prioritize understanding how best to implement it for long-term benefits. (Source: Yahoo)

JPMorgan Chase has maintained its top position in AI adoption within the banking sector, according to the Evident AI Index. This ranking, based on factors like talent development, innovation, and leadership, highlights JPMorgan’s dominance, especially in AI workforce size—employing more AI researchers than the next seven banks combined. The bank is also one of the few financial institutions reporting a return on investment (ROI) from its AI use cases.

Capital One ranked second in AI adoption, making significant strides by enhancing its AI skills and development capabilities. It leads in AI talent density relative to headcount and holds 38% of AI patents filed by the 50 banks analyzed. (Source: CIODive)

New York's financial regulator, the Department of Financial Services (NYDFS), has issued new guidance focusing on the cybersecurity risks associated with artificial intelligence (AI). This 11-page document encourages regulated entities to monitor and assess risks from AI tools, including threats from social engineering, cyberattacks, and the theft of nonpublic information.

While the guidance does not impose new requirements, it builds on existing cybersecurity regulations and emphasizes the importance of understanding AI-related risks. NYDFS Superintendent Adrienne Harris highlighted the necessity for financial firms to integrate AI into their risk management frameworks and to engage with stakeholders to understand the evolving technology.

The guidance also calls for multiple layers of security controls to ensure protection against cyber threats, including risk assessments, due diligence on third-party vendors, cybersecurity training, and effective data management.

This initiative follows California's recent veto of an AI safety bill, which raised broader regulatory questions about whether oversight should focus on AI model creators or applications. Despite the veto, many industry leaders anticipate that AI regulation will evolve, addressing the complexities surrounding its use and implementation. (Source: WSJ)

Perplexity AI, a leading generative AI startup, is launching several new services aimed at enterprise customers to enhance financial decision-making. Key offerings include an "Internal Knowledge Search" that allows organizations to simultaneously search internal files and the web, enabling more comprehensive research for tasks like due diligence.

The new "Perplexity for Finance" service provides real-time stock quotes, historical earnings reports, and detailed financial analyses, aiming to simplify decision-making in high-stakes financial environments. The company plans to integrate third-party data from sources like Crunchbase and FactSet for Pro customers, with future expansions into other sectors such as legal and healthcare.

Perplexity AI's tools are being likened to a more accessible version of traditional finance platforms like Bloomberg Terminal, potentially transforming capital market research. Competitors, including JPMorgan and Bloomberg, are also developing generative AI tools for financial analysis and client decision-making, posing a challenge to Perplexity. (Source: AnalyticsIndiaMag)

Retail

Google has revamped its shopping platform, integrating advanced AI to create a more personalized and efficient experience. The new Google Shopping, launching in the U.S., leverages AI and the Gemini models to enhance product discovery and research for its 45 billion product listings. The platform now offers AI-generated briefs that summarize key information for searches, helping users quickly find the best products. Features include dynamic filters for personalized results, a virtual try-on tool powered by AR, and a customized homepage feed that tracks your preferences and shopping history.

Shoppers can easily find the best deals with tools for price comparison, insights, and tracking, and a new personalized "Deals" page. Google also emphasizes the experimental nature of its AI-generated briefs, encouraging feedback to improve accuracy. The new shopping platform is accessible via the Google Shopping tab or shopping.google.com and will be rolled out across the U.S. in the coming weeks. (Source: Google)

Recent advancements in AI-assisted editing platforms are poised to significantly enhance productivity for eCommerce teams, facilitating faster creation of multilingual product descriptions and marketing copy. OpenAI has introduced Canvas, a new interface for ChatGPT that promotes collaborative writing and coding by allowing users to work alongside the AI in real time. This shift from traditional dialogue-based AI to interactive collaboration aims to create a seamless partnership between humans and AI, making the creative process more efficient.

Canvas includes features that streamline content creation, such as shortcuts for editing, adjusting text length, and enhancing code. These tools are especially beneficial for eCommerce businesses managing large product catalogs across various platforms and languages. Other companies, like Anthropic, are also exploring collaborative AI with updates to Claude that enhance team productivity through shared knowledge bases and customized responses.

The integration of AI in content creation workflows is transforming eCommerce operations. For instance, AI can rapidly generate initial product descriptions, allowing copywriters to focus on refining and adding persuasive elements. It can also suggest SEO-friendly phrases in real time, improving search rankings.

The benefits extend to quick revisions, as tools like Opus Pro and Descript enable faster edits, saving time and allowing teams to concentrate on creativity and strategy. AI-powered collaborative tools also enhance consistency across organizational messaging, making it easier for teams to align styles across various projects.

Additionally, AI can assist brands in testing products in different settings, like displaying sunscreen in various environments, thus catering to diverse markets while maintaining a personal touch. Overall, these advancements in AI editing tools are helping eCommerce teams produce higher volumes of content more efficiently while maintaining quality.(Source: PYMNTS)

Travel

Tesla has unveiled its long-awaited Cybercab, an autonomous robot taxi set to revolutionize urban transportation with a projected price of under $30,000 and an operating cost of $0.40 per mile. The vehicle was showcased at Tesla’s “We, Robot” event, which featured various themed neighborhoods highlighting the company's future vision. Production is slated to begin by 2027, but experts express skepticism about the timeline and technology readiness, particularly regarding regulatory hurdles.

The Cybercab is designed without manual controls, making it illegal to operate on roads in many regions. Industry experts worry about its regulatory prospects, noting that it lacks a framework for deployment in most countries. Despite these concerns, the introduction of Tesla’s Cybercab and a larger autonomous vehicle, the Robovan, could disrupt the transportation and logistics sectors by providing cost-effective ride-hailing and last-mile delivery services. They may also reduce congestion and emissions, offering benefits for those unable to drive, such as the elderly and disabled.

Visually, the Cybercab features painted aluminum body panels and lacks a steering wheel or pedals, emphasizing its fully autonomous nature. Inside, it accommodates two adults with a large center display and minimalistic design. Tesla's new AI system is designed to handle various driving scenarios and claims to make autonomous vehicles significantly safer than human-driven cars.

Additionally, Tesla plans to establish “Cybercab Hubs” for charging and cleaning the vehicles, utilizing inductive charging and robotic systems. The Robovan aims to expand Tesla’s market presence by serving as a flexible transport solution for people and goods.(Source: Pymnts)

Germany has launched its first AI-generated travel influencer, named “Emma,” developed by the German National Tourist Board to inspire potential visitors. Emma is designed to be an authentic and trustworthy brand ambassador, featuring a modern look with references to German culture. She will share travel tips, videos, and images on her Instagram account (@EmmaTravelsGermany) and engage with users through text conversations on the tourism board’s website. Emma can communicate in over 20 languages and is powered by advanced AI technologies.

The initiative aims to enhance the digital strategy of the tourism board, serving as a bridge between travelers and experiences in Germany. Long-term, Emma is expected to evolve into a personal travel companion, offering tailored itineraries based on individual preferences. The board envisions partnerships with hotels, airlines, and travel agencies to provide exclusive deals and enhance her role in promoting German tourism.

However, early interactions with Emma revealed basic responses, prompting the tourism board to emphasize ongoing improvements through machine learning, allowing her to respond empathetically to users’ needs.

Some social media users expressed concern that AI influencers like Emma could replace human content creators. The German National Tourist Board clarified that Emma is meant to complement human influencers, who foster genuine connections with audiences.

This move reflects a broader trend in destination marketing, as organizations increasingly invest in generative AI to provide personalized travel recommendations instead of generic guides. For instance, Brand USA recently appointed its first chief AI officer to drive similar initiatives in American tourism. (Source: SKIFT)

Software

Mira Murati, the former chief technology officer at OpenAI, is in the early stages of raising funds for a new AI startup, seeking over $100 million from venture capitalists to develop AI products based on proprietary models. While it’s unclear if she will take on the CEO role, Murati has reportedly been recruiting former OpenAI employees to join her venture.

Murati's departure from OpenAI in late September followed a tenure marked by significant contributions to projects like ChatGPT and DALL-E, and her involvement in a major partnership with Microsoft. She was a prominent public figure alongside OpenAI CEO Sam Altman and briefly served as interim CEO when Altman was ousted by the nonprofit board.

The new venture comes as Murati joins a trend of former OpenAI executives launching their own startups, with notable examples including Anthropic and Safe Superintelligence. (Source: Reuters)

Meta announced the release of new AI models from its research division, including a "Self-Taught Evaluator" designed to reduce human involvement in AI development. This model employs a "chain of thought" technique, similar to OpenAI's o1 models, breaking complex problems into smaller, logical steps to enhance accuracy in subjects like science, coding, and math.

Notably, the Self-Taught Evaluator was trained using entirely AI-generated data, eliminating the need for human input during this stage. This advancement points towards the potential for autonomous AI agents capable of learning from their own mistakes, a goal shared by many in the AI field. Such self-improving models could streamline the current, often costly process of Reinforcement Learning from Human Feedback, which relies heavily on human annotators.

Jason Weston, one of the researchers behind the project, emphasized that as AI capabilities grow, these models could surpass average human performance in self-evaluation. While other companies like Google and Anthropic have explored similar concepts of Reinforcement Learning from AI Feedback (RLAIF), they generally do not make their models publicly available, setting Meta apart in its approach. (Source: Reuters)

Hope you enjoyed this weeks Neurians AI Digest! Till next week…