On October 27, truck driver Mr. Zhang from Langfang, Hebei, accepted an order on the Huolala platform. Initially unaware of the cargo, he uploaded loading photos only to discover it contained fireworks. The platform’s AI safety system immediately flagged the shipment as hazardous. After communicating with the user, Mr. Zhang canceled the order, and the AI accountability system automatically analyzed the data, ruling him fault-free.
For gig economy workers like Mr. Zhang, avoiding algorithmic penalties has long been a challenge. To better protect their rights, the All-China Federation of Trade Unions has spearheaded "algorithm negotiations" with 15 platform companies, prioritizing workers' reasonable demands. Thanks to this initiative, Mr. Zhang and others no longer face penalties in similar scenarios. So far, 14 platforms have completed these negotiations, benefiting over 20 million gig workers.
**AI Algorithms Enhance Safety and Efficiency** "When I started hauling in 2019, canceling orders meant penalties—submitting photos, videos, and waiting for manual review. Now, the system auto-detects hazardous shipments like fireworks or gas tanks and absolves drivers of blame," Mr. Zhang explained.
On November 4, Huolala released its *AI Safety and Accountability Algorithm*, designed to improve driver safety. The system monitors orders for hazards like dangerous goods, unauthorized passengers, fatigue driving, or overloads, intervening with cancellations or alerts. Post-cancellation, the AI analyzes data against platform rules to assign accountability, often exonerating drivers for violations like prohibited cargo. Since implementation, hazardous shipments and unauthorized passenger cases have dropped 30% daily.
"The new algorithm shifts safety management from reactive to proactive, identifying risks preemptively," said Li Zhenying, Huolala’s safety operations head. "This was a key demand during algorithm negotiations with drivers. We’ll keep refining the system based on their feedback."
**Automated Exemptions for Delivery Riders** "I’ve seen riders break down over late orders—before, any delay meant penalties. Now, if restaurants are backlogged, the AI auto-files exemption requests," said Liu Xiaodong, an Ele.me team leader in Chengdu.
Liu attended Ele.me’s 2025 Algorithm and Labor Rules Negotiation Forum in September, where 10 riders nationwide collaborated with management in Shanghai. The resulting *2025 Ele.me Algorithm and Labor Rules Agreement*—the first of its kind—addresses pay, breaks, safety, benefits, and dispute resolution. It prioritizes dynamic factors like traffic controls and weather in dispatch algorithms, reducing delays and risks. Real-time traffic light integration in navigation aids efficiency, while penalties for lateness are phased out in favor of incentives.
Ele.me’s Party Secretary Xiao Shuixian noted post-Shanghai forums in nine provinces to gather rider feedback. "We’re upgrading fatigue prevention and ‘forced logout’ features to protect workers’ well-being," Xiao added.
**Timeout Waivers and Anti-Fatigue Measures** "Platforms vary widely—Meituan’s rules are too strict," riders voiced during a Beijing forum with seven Meituan representatives.
Meituan’s safety lead Xun Bin emphasized long-term safety goals, now managed via tiered interventions: alerts, pauses, or training. After nationwide negotiations, Meituan released its *2025 Rider Labor Protection Agreement*, piloting "timeout waivers" in 30+ cities. By year-end, the policy will replace penalties with points and training.
To curb overwork, Meituan’s "fatigue prevention" feature logs riders out after 12 hours, with hourly reminders post-8 hours. While welcomed, riders seek more flexibility.
Zhang Chenggang of Capital University of Economics and Business suggested improving restaurant ETA accuracy and traffic recognition to ease pressure. Initiatives like "red-light rewards" could promote safer deliveries, alongside dedicated lanes for riders.