Consulting ChatGPT?: Don’t talk to strangers


PARENTS used to remind their children never to talk to strangers as untold harm could befall them if they do. But in this day and age, the freedom found on the internet has allowed people to transcend time and space and throw reckless abandon on good parental advice.

The wonder is that the internet now allows people to not only talk to those they know but also people they don’t know. And in some cases, these “people” may not even exist.

Such was the experience of talking to artificial intelligence (AI) chatbot ChatGPT, which has made headlines for its novelty and, to a certain extent, notoriety. Teachers everywhere have raised alarm bells over papers being written not by the students but by their computers. Being able to tell the difference is proving to be a challenge.

This became the impetus for an experiment. So just like Alice in Wonderland, this reporter decided to follow the white rabbit wearing a waistcoat down the rabbit hole and asked ChatGPT a number of questions. Foremost of these questions was an economic one.

The Philippines has been toiling for years to find the right answers to solve development problems in order to uplift the lives of millions of Filipinos. Thus, we asked ChatGPT the question “How can we fight poverty in the Philippines?”

Our “new friend” gave a long answer that included seven key points such as access to education; job creation; social protection; women empowerment; infrastructure development; agricultural development; and addressing corruption.

These points, ChatGPT said, are crucial in fighting poverty in the Philippines, which it called a “complex and multifaceted issue” that cannot be addressed by just one solution.

“Fighting poverty in the Philippines requires a long-term and sustained effort from all stakeholders, including government, private sector, civil society organizations, and individuals,” ChatGPT said.

“By working together and implementing a comprehensive approach, we can address the root causes of poverty and create a more equitable and prosperous society,” it added.

Academician reacts

If anything, it’s as if a seasoned academician wrote the response. De La Salle University economist Maria Ella Oplas said she agreed with ChatGPT that poverty is indeed a complex and multifaceted issue.

Oplas, however, said prioritization is needed to fight poverty in the Philippines. She said attracting investments by investing in infrastructure should be a priority. This will reduce the cost of doing business in the country and create more jobs.

She added that in terms of agriculture development, it is not enough to talk about food security and rural development but also consider the fact that farmers remain a marginalized group in the country. This prevents the further development of the farm sector.

The economist added that it is not enough that access to education be improved; the feminization of education where boys are being left out, should also be addressed.

“We will not be able to compete locally and internationally with poor, uneducated human resources. Poverty results in uneducated children and vice-versa,” Oplas said. “Women empowerment and corruption are cross-cutting themes across all three issues.”

Canlas, Lanzona unimpressed

Other local economists were also unimpressed. Former Socioeconomic Planning Secretary Dante Canlas said the responses of ChatGPT to the question was “standard” and could easily be obtained from the website of the United Nations Sustainable Development Goals (SDGs).

Canlas said many of the recommendations have already been identified as anti-poverty measures for the past 30 years. These lessons have guided the principles behind the SDGs.

The SDGs aim to “eliminate poverty everywhere,”  through, for example, access to education, job creation, women empowerment, and social protection programs. There’s also access to basic services like clean potable water, and climate-change mitigation.

“The replies you got from ChatGPT are standard,” Canlas told the BusinessMirror via email. “The SDGs are comprehensive and they include the ChatGPT answers.”

Ateneo de Manila University’s Leonardo A. Lanzona Jr. agreed with Canlas and said the responses of ChatGPT were merely a summary of the ingredients of poverty elimination.

He raised three main concerns regarding ChatGPT and said that while the AI was able to describe and predict phenomena, an article coauthored by linguist Noam Chomsky in The New York Times noted that the AI fails to provide an explanation of or establish causal relationships.

“The machine can utilize academic findings that indicate causal relationships but, by virtue of its programming,  is indifferent as to how these findings have been generated,” Lanzona said.

Dangerous basis for policy

“The point is crucial because policy must be based on causal relationships, not just correlations. As economists, we struggle with these causal findings with the acceptance that we can be corrected and an openness to revise our findings. ChatGPT does not have these qualities, thus making it dangerous to use this as a basis for policy,” he explained.

Lanzona, like Oplas, also noted that while the laundry list of what can be done to address poverty in the Philippines is comprehensive, the question of prioritization comes into focus.

He said prioritization becomes crucial especially in light of limited resources. Following blindly the advice of AI such as ChatGPT could mean running the risk of not having sufficient resources to back up the plans and programs needed.

The value of having a human touch in policymaking is part of what will drive projects forward. Knowing his limitations, Lanzona said, would place a person in a better position to determine constraints and other problems that could arise from policymaking.

One such limitation that ChatGPT may not be able to understand is the extent of the effects of Covid-19 on the country and the economy, as not much has been written about the “full effects” of the pandemic.

“ChatGPT has limited knowledge of events that transpired after 2021. As such, it does not have enough data on the full effects of Covid-19, which continues to devastate the country. It is rather alarming that health and post-pandemic issues such as learning losses did not become one of the factors needed in fighting poverty,” Lanzona said.

With these inputs from real economists, it may be safe to say that ChatGPT is not the great economist that our country needs to solve deep-seated problems such as poverty.

And so, unlike Alice, the end of the rabbit hole may be a dud. This BusinessMirror reporter could have done better than to ignore good parental advice. Indeed, don’t talk to strangers.

Image credits: Pingingz |