CCTV has uncovered disorderly practices among volunteer-filling agencies, including “hiring actors to fake popularity and using AI to generate quick answers during consultations”. Nandu has identified multiple issues with AI-based volunteer-filling tools.

After the gaokao results are announced, selecting colleges becomes a top priority for countless families. Once the scores are in, the stress of making the right choices truly sets in.

Capitalizing on this anxiety, various AI-powered tools have flooded the market, promising one-click plan generation using large models and big data. For many parents, these tools have become a beacon of hope.

According to Southern Metropolis Daily, reporters from the Nandu Big Data Research Institute, from a user's perspective, tested several apps and mini-programs that claim to use AI to assist in gaokao volunteer selection. The reporters, posing as a candidate with a 640 in physics in Guangdong Province and an interest in computer science, mainly tested multiple mainstream products. The findings revealed that the issues were far more severe than anticipated.

The ‘Diebian Volunteer’ app listed the electronic information category at Peking University as a top ‘stretch goal’, showing a 41% chance of admission.

However, according to the 2025 Guangdong one-point-one-rank table, a score of 640 places the candidate at the 6,295th rank in the physics category, while Peking University’s admission rank in the Guangdong physics group is around the 105th rank. The discrepancy of over six thousand ranks is far beyond a reasonable stretch, making the recommendation seem like a digital fantasy.

Distorted basic data is even more prevalent. Among the stable colleges recommended by the ‘Xing Volunteer’ mini-program, the computer science category at South China University of Technology is marked with a minimum admission score of 616 and a rank of 13,157 in Guangdong in 2025. Yet, the official website of South China University of Technology indicates that the actual minimum score for this major is 635 and the rank is 8,015, showing a significant deviation. Even the true rank corresponding to a score of 616 is 16,493, revealing inconsistencies in the platform's data.

Several apps, including ‘Gaokao Direct Train’, ‘Sunshine Gaokao Volunteer Filling Guide’, and ‘Reliable AI’, all exhibit significant deviations in admission scores and ranks, with errors exceeding two hundred ranks for some colleges. User reviews frequently complain about ‘outdated data’, ‘inconsistencies with the official website’, and ‘untimely updates’. More notably, these products, which claim to offer ‘precise filling’, almost never prominently remind users that ‘results are for reference only and official data shall prevail’.

While data accuracy is a concern, the monetization strategies are clear. Almost all tools employ a ‘free trial + membership unlock’ model, using basic functions to attract users and then charging for complete plans and in-depth analyses.

The ‘Jiashu AI Volunteer’ mini-program only displays a limited number of colleges. To view the full recommendations, users must purchase a VIP membership, with prices ranging from tens to hundreds of yuan. The page also uses phrases like ‘only one step away from your ideal college’ to heighten anxiety.

Some platforms use AI as a lure to sell high-priced one-on-one consultations. In ‘Xing Volunteer’, a 98.5 yuan expert card claims to outperform most gaokao volunteer-filling experts in a day, and a 399 yuan membership can be deducted from offline consultations, with the final one-on-one service priced mostly between 3,000-5,000 yuan.

CCTV News investigations have exposed the chaotic practices of offline volunteer agencies charging exorbitant fees: consulting service prices on the market vary widely, ranging from hundreds to tens of thousands of yuan.

Several large institutions categorize their fees into three tiers based on so-called ‘teacher qualifications’, priced at 4,980 yuan, 8,980 yuan, and 12,980 yuan respectively. Many institutions promote ‘AI + human double screening’ to justify their high service fees. Many parents admit that they pay to bridge information gaps and seek peace of mind, but they remain skeptical about the actual effectiveness of the services.

The consultants displayed on the platforms mostly hold titles of senior career planners, but there is no officially recognized qualification for volunteer planners in China, and these titles are mostly self-proclaimed by the institutions.

An insider interviewed by CCTV revealed that these roles are essentially sales positions, with most employees having less than two months of experience and no practical experience in volunteer filling. The pre-job training at these institutions focuses on fabricating work histories and using photo editing to create expert personas, with sales personnel posing as professional teachers to interact with parents.

The institutions also drive traffic from live streams to WeChat groups, arrange ‘actors’ to create a false impression of busy consultations, and then promote services based on parents' willingness to pay, deliberately exaggerating risks and creating panic before the filing deadline to rush orders. Some institutions even have senior planners who rely entirely on AI searches to find answers temporarily during consultations and cannot provide personalized advice.

In response to these phenomena, experts and officials have issued multiple warnings. Chen Hao, an associate professor at the School of Artificial Intelligence at Xidian University, pointed out that the core of AI-based filling is information organization and cannot replace human decision-making. Parents must cross-check with official data.

Li Qing, the director of the Admissions Office at Beijing University of Chemical Technology, also stated that data lag and algorithmic limitations can easily lead to deviations, and recommended relying on information released by college official websites and examination authorities.

The Guangdong Provincial Education Examination Authority also reminds students to fill out their volunteer forms through official channels and to be cautious of third parties leaking personal information and various enrollment scams.

Stripping away the AI veneer, the core of this business is not new. From the early days of filling manuals and offline lectures to the current AI plans, only the format has changed; what remains unchanged is the exploitation of gaokao anxiety. Parents are willing to pay essentially to buy certainty.

However, many AI tools cannot even guarantee basic data accuracy. Their so-called intelligent algorithms are nothing more than applying templates of past data and adding large model rhetoric.

AI is not without value. It can quickly screen college ranges, summarize professional information, and help parents save time. However, treating it as a standard answer and directly copying it is entrusting twelve years of hard work to a program that will not take responsibility.

There is no one-size-fits-all solution for gaokao volunteer filling, nor is there a perfect plan that can be generated with one click. Those tools that claim to ‘squeeze every point’ often first squeeze the wallets of parents.

There are no shortcuts to filling out volunteer forms. Verifying official data, weighing interests and employment, and considering cities and colleges all require careful steps taken by oneself. After all, crucial life choices cannot be entrusted to an AI that cannot even calculate ranks correctly.