This story hovered under the radar. It deserved to be there only if we ignore the disturbing implications.
Regional Trial Court 53 in Pangasinan overturned the results produced by the Automated Counting Machine (ACM) in the 2025 Rosales vice mayoral race. The ruling upheld the results of the manual count, which contradicted the machine count.
At stake is a minor post that many might think is not worth the headline. But the real story here is that the much-touted automated count is, indeed, fallible.
If the automated count can fail in a small town, it can fail elsewhere. If it can fail in a local poll, it can fail in a national count.
We have long been comfortable in the thought that the ACM is infallible. It is a more comforting thought, after all. I have not monitored any statement from the Comelec regarding the Rosales ruling. This makes the event even more discomforting.
Before this ruling, the Comelec kept reassuring us that the automated counting system produced near-perfect results. This is true only if we compare the machine results with the random manual counts. As the Rosales case illustrates, some results may fall through the cracks of a random check.
The machine count reversal in Rosales had its own costs by way of contested results, delayed project approvals, interruptions in the procurement process and curtailment of public services. Should contested results happen on a larger scale, the disruptions become proportionally larger.
More important, the overturn of the Rosales results casts unwanted doubt on the reliability of machine counts. The importance of credibility in election counts cannot be understated.
In 2028, the nation faces what will likely be a polarizing election contest. Rivalry will be intense. Passions will be high. Public trust in the counting system we have adopted will mean everything.
We cannot afford dissipation of public trust in the electoral count. The fallout from a widely contested count will not only paralyze government. It could spell a meltdown in our electoral democracy.
The Comelec ought to assure us that the Rosales case is a truly isolated instance. But to do that, the poll body must provide us convincing forensics.
Beyond that, more effort should be put to conserve the audit trails and the electronic safeguards. No stone must be left unturned. The 2028 count must be believed by all voters.
We cannot possibly overstate the horrors a failed count might bring.
Disemployment
The dire warnings about the effects of widening use of artificial intelligence (AI) on employment levels are based on sound evidence. Large companies are restructuring or merging in the face of new technological realities.
Giant corporations have been reducing the number of workers they employ. In 2022 and 2023, Amazon eliminated 27,000 jobs. During the same period, Meta cut 21,000 jobs. Accenture, which employs 50,000 Filipinos, shed 22,000 positions globally in 2025 while investing $923 million on severance and retraining half a million surviving employees in generative AI.
Telstra, Australia’s largest telco, laid off 3,500 workers over the past three years. The company announced yet another round of layoffs tied to its AI joint venture with Accenture this month.
These are blue chip corporations. They are not laying off workers because they could not afford to retain them. They are trimming down jobs because it makes better business sense to replace human labor in routine and repetitive functions with automated systems. Better efficiencies justify the layoffs.
The wave of technologically induced disemployment will not spare the vibrant economies of Southeast Asia. Governments in the region must move with haste in confronting the new challenge posed. This will be a matter of great concern in the various ASEAN meetings the Philippines is hosting this year.
Government responses vary widely across the region. Singapore, as usual, moved earliest and most deliberately. Under its Skills Future Level-Up program, every Singaporean citizen turning 40 automatically receives S$4,000 in non-expiring training credit. This can be used across a curated list of over 7,000 courses tied to actual hiring outcomes.
Workers aged 40 and above qualify for up to 90 percent subsidy on course fees. Those who opt for full-time long-form retraining receive monthly allowances of up to S$3,000, capped at 24 months.
The program recognizes that mid-career workers face higher obsolescence risk and higher opportunity costs for retraining. By May 2025, over 36,000 Singaporeans availed of the credit. Like their government, they understand that they either retrain or be replaced.
Thailand and Indonesia have been undergoing a consolidation wave. Large corporations have merged to improve their efficiencies. In the process, merging corporations have offered structured exits for employees aged 45 or more. In both economies, workforce transformation programs have been largely sector-specific and voluntary.
Malaysia, for its part, established its Human Resources Development Fund and the HRD Corporation. This set-up levies a mandatory training contribution from employers and redistributes this as grants for worker upskilling.
Vietnam has been slow in building a policy response to the employment challenge. Reflecting this, Vietnamese workers have the highest AI anxiety rate in the region.
The Philippines’ response to the disemployment challenge is small compared to the possible magnitude of the disemployment challenge. The 2025 budget allocated P70 million for TESDA to establish a labor market information system, establish an AI-powered course and start an IoT training system.