AI News Digest 2024-06-28
AI MaaS expands, AI adopted in eyecare, NBC uses AI for voice, payment fraud targeted, fairness gaps in AI medical imaging.
Productization Trend Extends to AI Through MaaS
Productized AI Models Drive 'En Masse' Intelligence Forbes
The productization trend has extended to artificial intelligence (AI) through Model-as-a-Service (MaaS), making prebuilt, preconfigured ML models more accessible. NTT Data launched its Tsuzumi large language model through Microsoft Azure AI MaaS service, enhancing scalability and adaptability. SAS is also offering industry-specific AI models for various sectors, democratizing AI by making it usable and workable in different environments.
Adoption of Artificial Intelligence in Eyecare Industry
Optical Gains Traction in Adopting Artificial Intelligence Vision Monday
Optical companies and eye care professionals (ECPs) are increasingly implementing artificial intelligence (AI) in their business models to enhance productivity and provide better services. Despite concerns about reliability and security of AI systems, a growing number of optical businesses are seeing positive results. Examples of AI implementation include targeted marketing campaigns, patient education, and improving productivity for tasks like charting and data analysis. Additionally, AI is being used by eyewear companies and contact lens manufacturers for creative campaigns, personalized lens designs, and enhancing efficiencies in supply chain management. Companies are cautious about privacy and security but are embracing AI to provide superior customer experiences. Contact lens companies such as Johnson & Johnson Vision and CooperVision are leveraging AI for supply chain logistics, patient screening and diagnosis, and scientific research. Lens manufacturers like EssilorLuxottica, Shamir, Zeiss, and Rodenstock are using AI for personalized lens designs and to gather wearer data for improving visual experiences. These applications indicate a shift towards AI adoption in the eyecare industry.
NBC Utilizes Artificial Intelligence to Recreate Al Michaels' Voice for 2024 Paris Olympics
NBC Utilizes Artificial Intelligence to Recreate Al Michaels' Voice for 2024 Paris Olympics The New York Sun
NBC plans to use artificial intelligence to recreate Al Michaels' voice for the 2024 Paris Olympics, providing customized Olympic recaps for Peacock subscribers during the event. The AI-generated voice will greet users by name, highlight their specific sports interests, and offer customized 10-minute highlight packages. A team of NBC Sports editors will review the AI-generated content for accuracy and context.
Payment Fraud Prevention Through Generative AI Tools
Payment Fraud Prevention Through Generative AI Tools PYMNTS.com
Generative AI tools are gaining traction in payment fraud prevention. Traditional rules-based systems are becoming inadequate to deal with the complex fraud schemes, leading to high false positives and limited adaptability. Predictive AI and generative AI, using unsupervised or semi-supervised learning techniques, offer improved capabilities in detecting fraud patterns in unstructured data. Generative AI enables continuous learning and can produce synthetic datasets for model training without privacy concerns. Leading companies like Visa and Mastercard are actively using generative AI for payment fraud detection. The technology shows potential in reducing false positives, improving accuracy, efficiency, and cost savings. However, its adoption faces challenges such as privacy concerns, bias, and regulatory hurdles.
Fairness gaps in AI medical imaging models
Study reveals why AI models that analyze medical images can be biased MIT News
A study by MIT researchers reveals that artificial intelligence models used in medical diagnoses tend to have fairness gaps and may use demographic shortcuts, resulting in discrepancies across different demographic groups. Strategies to debias the models have shown some success but may not necessarily hold when applied to new patient groups from different hospital datasets.