What the Numbers Show About AI’s Harms
What the Numbers Show About AI’s Harms
https://time.com/7346091/ai-harm-risk/
Publish Date: 2026-01-19 09:05:00
Source Domain: time.com
Certainly! Here are five key points from the article, presented as an unordered list:
- Rising AI Incidents: There has been a substantial increase in reported AI-related incidents over the past couple of years, with a 50% rise from 2022 to 2024, further escalating in the first ten months of 2025.
- Harm and Limited Reporting: AI is increasingly causing real harm, but data on incidents is limited due to the difficulty in capturing all media reports and incidents. Only the most critical issues are typically reported to authorities.
- Deepfake Scams and Data Classification: The number of AI-generated deepfake scams has skyrocketed, while other types of incidents like misinformation and discrimination have declined. Efforts to classify incidents are underway using both crowdsourcing and AI tools to detect patterns and trends.
- Challenges and Legal Frameworks: The AI Incident Database faces challenges due to the lack of a structured approach, which makes trend analysis difficult. Regulatory frameworks such as the EU AI Act and California’s SB 53 require reporting of serious incidents but do not capture the full scope of AI harms.
- Attention to Emerging Risks: While immediate harms from scams and manipulated content are evident, longer-term societal harms like privacy issues, misinformation, and erosion of rights also need attention, underscoring the need for comprehensive and evolving regulations.
