AI expert Jeff Heaton presents his quarterly update of recent technological advances and outlines their relevance to the insurance industry. In this installment: a discussion of new launches from top GenAI companies and predictions for 2025.
Have a question or comment for Jeff? Join the conversation.
As we begin 2025, many exciting developments in the field of Generative AI (GenAI) have emerged. This update highlights significant advances from top GenAI companies and their impact on the insurance industry and concludes with some predictions for 2025 and beyond.
Anthropic Claude 3.5’s new "computer use" capability represents an evolution in automation comparable to robotic process automation (RPA). Like RPA, it enables interaction with computer systems to automate repetitive tasks. However, while RPA requires pre-defined workflows and specific programming to perform actions, Claude’s GenAI capabilities allow it to adapt dynamically to different contexts. This flexibility means Claude can handle complex tasks involving exceptions or variable inputs, making it especially valuable in the insurance industry, where workflows often involve diverse systems and data formats.
For insurers, this capability complements and extends beyond traditional RPA. For example, RPA might automate claims entry by following a rigid script, but Claude can go a step further – navigating unstructured portals, responding to unexpected errors, and consolidating data from multiple systems in real time. It also can assist underwriters by gathering applicant data from various online sources, checking policy terms in legacy systems, and even ensuring compliance by summarizing regulatory updates across platforms.
While RPA excels at streamlining structured, rule-based processes, Claude 3.5 brings flexibility and cognitive understanding to tasks that require adaptability. This capability is particularly relevant in insurance, where manual interventions are often needed for complex workflows. As GenAI continues to evolve, insurers have an opportunity to integrate it with existing RPA systems, creating a hybrid approach that combines the precision of RPA with the adaptive intelligence of AI to improve efficiency and decision-making.
OpenAI’s ChatGPT recently introduced a powerful new feature: the ability to search the web in real time and provide source citations. This update marks a significant leap in utility, combining GenAI capabilities with up-to-date, reliable information retrieval. Users can now receive not only comprehensive, AI-generated insights but also references to verify and further explore the information provided. This enhancement addresses one of the primary limitations of earlier iterations of AI models, which relied on static, often outdated knowledge.
For insurers, this search functionality brings several game-changing advantages.
One is staying informed about the latest developments in insurance regulations, market trends, and customer expectations. For example, an underwriter might quickly pull up the latest data on actuarial tables or changes in state-specific laws without leaving the ChatGPT interface.
Additionally, the ability to cite sources adds a crucial layer of transparency. Insurers can use ChatGPT to draft policy explanations, market analyses, or business presentations, knowing the data can be verified against authoritative sources. This builds trust both internally and with clients. For example, an insurer could use ChatGPT to provide real-time, evidence-backed answers to customer queries, boosting credibility and satisfaction.
Looking ahead, combining this search functionality with other insurance tools, such as claims management systems or fraud detection platforms, could further streamline workflows. By integrating real-time search and GenAI, insurers can both respond to the present and anticipate future challenges.
Google recently introduced its latest quantum computing breakthrough, a chip called Willow, which represents a significant step forward. Detailed in a recent article in Nature, Willow is a 105-qubit processor designed to push the boundaries of quantum computing. It demonstrated its power by solving a benchmark problem in under five minutes – a task that would take the fastest classical supercomputers a length of time far exceeding the age of the universe. This milestone highlights both the potential and the growing maturity of quantum computing technology.
Adding qubits to a quantum computer dramatically increases its power, allowing it to solve far more complex problems. However, qubits are fragile and prone to errors, which has limited their usefulness. Willow reduces these errors by combining groups of qubits into more stable units, making the system more reliable as it grows.
This breakthrough in chip technology brings quantum computing closer to practical applications in fields like medicine, finance, and energy.
The immense computational power of quantum computers could transform how we train and fine-tune AI models, leading to AI systems that are more accurate, faster to develop, and capable of handling unprecedented amounts of data. For insurers, this could mean faster risk analysis, more precise underwriting models, and the ability to predict and adapt to emerging trends with greater accuracy.
OpenAI has unveiled Sora, an AI-driven text-to-video model that enables users to generate videos through simple text prompts. This tool is accessible to ChatGPT Plus and Pro subscribers, allowing them to create videos up to 20 seconds in length with resolutions reaching 1080p. Features include video generation from text, image animation, and video remixing to deliver a versatile platform for creative expression.
To address potential misuse, OpenAI has implemented safety measures within Sora. The depiction of real people is restricted to a select group of invited testers to mitigate risks such as impersonation and the creation of deepfakes. Additionally, generated videos include watermarks and C2PA metadata to indicate AI involvement, ensuring transparency and accountability.
The launch of Sora represents a significant advancement in AI-generated content, positioning OpenAI at the forefront of consumer-accessible video generation technology. By integrating Sora into its suite of tools, OpenAI continues to expand the creative possibilities available to users, while maintaining a commitment to responsible AI deployment.
GenAI is expected to transform multiple industries, including insurance, in the years to come. Experts predict AI will enable personalized treatment plans in healthcare, streamlined risk assessment in finance, and tailored educational content for individual learners, just to name a few examples. Autonomous AI agents are expected to become integral to enterprise operations, performing tasks such as recruitment, marketing, and billing with minimal human intervention. This could reshape traditional software systems and drive operational efficiency across sectors.
However, this rapid integration will bring significant infrastructure and energy challenges. Global data center energy consumption is projected to double by 2030, fueled by power-intensive AI applications. Technical advancements in AI are also expected to become more complex as initial breakthroughs plateau.
Alongside these technical hurdles, ethical and regulatory considerations will take center stage. Robust AI governance frameworks will be essential to address issues like bias and transparency, with governments and companies working to balance innovation with regulatory compliance.
AI-driven productivity gains will contribute to economic growth, while simultaneously prompting workforce transformation. Companies will need to invest in reskilling and upskilling employees, focusing on emerging roles like prompt engineering and data literacy. As AI continues to evolve, it will necessitate thoughtful planning to harness its potential responsibly while mitigating risks associated with its widespread adoption.
While it is difficult to predict exactly how far this evolution will advance in 2025, it is clear thoughtful progress in all areas must proceed quickly to keep pace with technology. It will undoubtedly be another exciting year in GenAI, and I look forward to bringing you more updates and their implications for insurers as we move through it.
Have question or comment for Jeff? Join the conversation.
