Recently in mainland China, a journalist embarked on a day trip to Beijing with the assistance of generative AI tools, only to find themselves repeatedly let down. Despite the promise of a well-planned itinerary, they struggled to locate recommended breakfast and lunch spots, while the suggested cable car service had also ceased operations. Ultimately, hunger and cold weather forced a premature end to the trip.
On November 1, a mainland media outlet published an account detailing the journalist’s experience, which began with an intent to explore Beijing in the heart of autumn. The plan was to utilize four different generic AI models to gather travel recommendations and evaluate them through a firsthand experience.
According to the generated itinerary, the journalist was supposed to start their day with breakfast at a snack shop near the North Plaza of Beijing West Station, before heading off to Xiangshan (Fragrant Hills). However, after a thorough search and some cellphone navigation, it became clear that the suggested location did not exist, and the nearest alternative branch was over a fifty-minute walk away. Faced with the dwindling time, the reporter had no choice but to forgo breakfast and venture out on an empty stomach.
Upon arrival at Xiangshan, the plan hit another snag when the recommended cable car was found to be out of service due to upgrades and renovations. Undeterred, the journalist opted to climb the mountain via a nearby trail, but after a short distance, they realized that they were running out of time to meet the scheduled departure for the next planned stop. Consequently, they could only take a brief look around the foot of the mountain.
The journalist then arrived punctually at the western gate of the Summer Palace, eager for lunch, but was again met with disappointment as the AI-recommended dining establishment turned out to be nonexistent. The closest available option was well over an hour’s walk away. At that point, with clothes and shoes drenched from the rain, the journalist felt cold and hungry. All they could do was admire the scenic spots from afar, leading them to conclude their journey prematurely.
In a subsequent round of testing, the journalist attempted to utilize a specialized AI developed by a tourism platform, yet the route still encountered issues. The recommended bus service failed to have a designated stop that matched the travel guide, necessitating a thirty-minute walk from the nearest bus station to the East Palace Gate of the Summer Palace. Alternatively, entering through the North Palace Gate would involve only a short walk of thirty to fifty meters, but adjustments in the planned itinerary would still be required.
Jin Zhun, Secretary-General of the Institute of Tourism Research at the Chinese Academy of Social Sciences, commented on this testing report, highlighting that the information provided by AI might lag behind, impeding timely reflection of changes in site availability and transportation regulations. He pointed out that the planning often presumes smooth transit connections and can overly schedule activities, making it challenging to adhere to the proposed timeline during actual outings.
Furthermore, Jin indicated that the specialized AI offered by the travel platform similarly suffered from lacking comprehensive information.
Another individual, an AI hardware product manager named Si Peng, who has been residing and working in Shanghai, shared his own experience of vacationing according to an AI-generated guide. He recounted, “Except for the departure time, everything else was unexpected.”
He elaborated on a trip from Wuhu in Anhui Province to Anji in Zhejiang Province, where traffic congestion during a holiday caused him to be stuck on the highway. Though the itinerary indicated lunch should be at 11:30 AM, he found himself still in traffic and only managed to arrive at his hotel by 6:30 PM, thus missing planned afternoon activities.
These issues are not isolated, as other tourists have expressed similar frustrations. A report from China News Network detailed how a photography enthusiast, Mr. Zhao, consulted AI for the closing time of a photography exhibition in Beijing before scheduling his visit, only to arrive and discover that the exhibition had already ended, resulting in over 1,000 yuan spent on round-trip high-speed train tickets.
Another traveler, named Da Lu, recounted in the same report that AI had informed her that a particular tea museum was free of charge. However, upon driving for an hour to reach the location, she learned that non-hotel guests were required to purchase admission tickets. Consequently, she and her companion ended up spending 136 yuan just to get in and enjoy two cups of white tea.
This series of unfortunate events underscores critical challenges faced by travelers relying on AI-generated travel advice in a rapidly changing landscape. As the intersection of technology and tourism continues to evolve, it raises important questions about the reliability and timeliness of information provided by these innovative tools.