The widespread adoption of Generative AI (GenAI) in applications offers substantial advantages but also introduces various threats because of the myriad components they comprise. To ensure the integrity of AI/ML systems, organizations should manage every component through an AI Bill of Materials (AIBOM) to inventory the data, models, and infrastructure used.
Developers, data scientists, and security experts should advance their AI maturity by adopting AIBOMs to secure and optimize their AI systems.
Retailers are facing a surge in shoplifting–highlighting a need for advanced techniques to deter theft effectively. AI tools offer promising results to deter shoplifters and allow retailers to not only reduce theft but also improve overall operational efficiency and customer experience. Tech leaders must learn about the latest AI methods that will enhance their store’s security.
State-of-the-art AI models require powerful hardware to run and they have enormous file sizes–leading to high hosting and inference costs. SMEs are unable to integrate these models into their applications because of their limited budget. Sparsity is a technique that prunes a model’s parameters leading to a smaller, faster model that can run on-device. An SME’s IT team can use sparsity to integrate powerful models into their applications while reducing hosting and inference costs.
It has become easier to create AI applications due to the ease of integration by using APIs. High cost is one challenge when frequent API calls are made to LLMs with similar content to add context. Prompt caching, or context caching, creates a cache to solve this challenge. AI engineers must use prompt caching to decrease inference fees and reduce latency.
High-quality data is an important requirement for a successful AI model. Meta’s Llama 2 and 3 were trained on 2 and 15 trillion tokens, respectively. One method of obtaining data is through web scraping. This has led to many businesses suing AI developers for using copyrighted material. IT leaders and Content Strategists can prevent their data from being scraped by blocking web crawlers that gather data for AI training.
Major device manufacturers are susceptible to a vulnerability called PKfail. PKfail allows bad actors to execute malicious firmware at boot time on devices. Cybersecurity professionals must examine their devices to remedy this vulnerability before it is exploited.
With the rise of e-commerce and growing customer expectations for fast, reliable deliveries, optimizing last-mile delivery has become a critical priority for businesses. This article offers Tech Leaders and Logistics & Supply Chain Managers key strategies to enhance last-mile delivery efficiency, reduce costs, and improve customer satisfaction.
AI has a significant positive impact on sports operations, efficiency, and viewer engagement. Many teams already use AI to increase their competitive edge. Now is the time for their sporting bodies to follow their example so that AI use becomes standard practice for their sporting events. Technology leaders must guide large sporting bodies to get started and become pioneers in the sporting industry.
Serverless computing provides ease of development, deployment, and management with features such as security, redundancy, scalability, and an affordable pricing model. This service can also lead to high unexpected costs if prevention methods are missing. IT professionals in SMEs must implement prevention methods to tame unexpected costs that can cripple the enterprise.
2024 saw significant shifts in technology, with the EU's AI Act and DMA impacting businesses alongside the rise of modular laptops and the persistent threat of cyber attacks. This review highlights some of the developments that interested IT leaders. This list suggests CIOs and IT executives should continue to prioritise compliance, evaluate new technologies, and strengthen cybersecurity in 2025.