Meta Plans to Shift 90% of Product Risk Assessments to AI, Raising Concerns

Meta's Shift to AI-Driven Risk Assessments

Meta, the parent company of Facebook, Instagram, and WhatsApp, is reportedly planning a significant change in how it evaluates potential risks associated with new features and updates across its platforms. According to internal documents obtained by NPR, the company aims to automate up to 90% of its risk assessments using artificial intelligence, replacing the current system that relies heavily on human reviewers 1

1

.

The New AI-Driven Process

Under the new system, product teams will submit questionnaires about their projects to an AI system, which will provide an "instant decision" identifying risk areas and requirements 2

2

. This automation is expected to streamline the decision-making process and allow for faster rollouts of new features and updates. The AI system will be responsible for evaluating various aspects, including:

1. Privacy concerns
2. Potential harm to minors
3. Content integrity (misinformation, violent content)
4. AI safety

Concerns and Criticisms

The shift towards AI-driven assessments has raised concerns among current and former Meta employees. Critics argue that this change could lead to overlooking serious risks that human reviewers might have identified . A former Meta executive, speaking anonymously, warned that reducing scrutiny could create higher risks and make it less likely to prevent negative externalities before they cause real-world problems 1

1

.

Meta's Response

Meta has defended the move, stating that human expertise will still be used for "novel and complex issues," while AI will handle "low-risk decisions" 2

2

. The company claims to have invested billions of dollars in supporting user privacy and aims to balance innovation with compliance 5

5

.

Implications for Different Regions

Interestingly, the internal documents suggest that users in the European Union may be somewhat insulated from these changes. Decision-making and oversight for products and user data in the EU will reportedly remain with Meta's European headquarters in Ireland, likely due to stricter regulations such as the Digital Services Act 1

1

.

Broader Context of AI Integration

This move towards AI-driven risk assessments is part of a larger trend at Meta to leverage artificial intelligence across its operations. Earlier this week, the company released its Integrity Reports for the first quarter of 2025, stating that it has begun using AI models to remove content from review queues in certain scenarios 4

4

.

As Meta continues to push for efficiency and faster product development, the balance between innovation and thorough risk assessment remains a critical point of discussion. The success of this AI-driven approach will likely be closely watched by both industry observers and regulatory bodies in the coming months.