What: Coresight reviews what retailers and companies need to know when considering moving forward with AI

Why it is important: missteps can prove catastrophically wrong in the realms of corporate finance, customer relationship or operations if AI is deployed without being aware of the involved risks.

The global generative AI market is estimated at $79.8 billion in 2024, growing to $235.5 billion by 2028.
Pre-deployment challenges include ensuring high-quality, unbiased training data free of proprietary information. Data poisoning and intellectual property concerns are also risks.
During deployment, models face knowledge gaps after training ends and can experience drift over time, requiring monitoring and potential retraining.
Models need supervision as they cannot determine if outputs are correct or appropriate. Risks include hallucinations, toxicity, biases, and data leaks.
Post-deployment, companies must comply with regulations and corporate policies, including audit trails and privacy protections. Cybersecurity is a key concern.
For retailers, early adopters defining use cases will gain advantages. Those interacting directly with models face higher risks. Ongoing monitoring for inappropriate content is crucial to protect brand image.
The technical requirements create opportunities for vendors offering model management, monitoring, governance and cybersecurity solutions.

Coresight: being aware of the challenges and risks of using AI