Generative artificial intelligence (AI) models like ChatGPT have exploded in popularity in recent months. These AI systems can generate synthetic text, images, audio and video that closely resembles human-created content. Major tech companies are racing to integrate generative AI into their platforms and services. But this rapid development and deployment raises serious concerns about potential consumer harm that must be addressed.

In a new report, the Norwegian Consumer Council highlights the consumer angle of generative AI. While these systems may provide some benefits, they also introduce vast risks. The report describes challenges relating to:

• Power concentration in the hands of big tech companies, who gain significant control over generative AI systems. This allows them to potentially entrench their dominant positions.
• Lack of transparency and accountability. Generative AI systems are often opaque “black boxes”, making it hard to audit them for errors, bias and other issues.
• Manipulation of consumers through techniques like deepfakes or anthropomorphised chatbots designed to influence emotions and opinions.
• Biases and discrimination that can be replicated or introduced, due to issues with training data or content moderation practices.
• Privacy violations, with personal data often used without proper consent to develop the models.
• Security vulnerabilities that could facilitate new types of fraud and cybercrime.
• Replacing human roles like moderators and creators with automated systems, which may reduce quality and exploit precarious labour.
• Environmental harms from the massive computing power required to develop and run these AI models.

To address these concerns, the report calls for better enforcement of existing regulations like consumer protection, data protection and product safety laws when applied to generative AI systems. It also recommends new safeguards like transparency requirements, impact assessments before deployment, and making developers responsible for harms.

As consumers start interacting with generative AI on a regular basis, it’s crucial that policymakers act swiftly to minimise risks and ensure these systems respect consumer rights. The report provides a timely overview of key issues and recommendations for oversight of this rapidly evolving technology.