Monday’s AI

It was supposed to be about artificial intelligence, but with the presence of Rep. Joey Salceda at the Monday Circle’s bi-weekly business forum, he quickly demonstrated that AI technology still has a lot to learn yet from intelligent humans.

Resource speaker Dominic “Doc” Ligot, a data analyst, researcher, software developer, entrepreneur and technologist who is an advocate for data literacy, AI ethics, data ethics and social impact from data, explained that AI tools are at the forefront of changing technology and that AI depends on the input of good data.

AI, unfortunately, Doc Ligot admitted, can be “polluted” by too much conflicting data, an example of which, he cited, is the Waze app.

It was interesting to learn that the majority users of the popular Waze app are actually motorcycle riders, thus, some of the alternative routes it recommends are narrow winding lanes that are not actually suitable for cars or vans, thereby confusing car and van owners who then become frustrated with the app.

Doc Ligot admitted that chatbots are also “polluted” because “everything has been put in them.” The response of the generative AI chatbots like ChatGPT, he said, would also depend on how the question is asked. A straightforward question would get a straightforward answer, but a malicious question would also likely get a malicious answer.

Additionally, AI tools such as ChatGPT, according to Doc Ligot, is a language-driven learning and informative tool that is not yet capable of doing mathematical calculation, following a query from Cong. Joey if it can do the Fibonacci sequence calculation. For mathematical calculations, Excel is still the tool to rely on, Ligot recommends. In the future, though, AI may eventually do mathematical calculations.

It was also informative to learn that Filipinos are still among those resistant to using AI even as a useful learning tool. A similar trend that has occurred in the transition phase from one technology to another, with Ligot citing the transition in art from manual painting or drawing to photography, from rote memorization in mathematics to the use of calculators and even the downshift from grammatical english writing to the so-called “jejemon” text writing that is now used with the modern cellular phones, and the use of Wikipedia as a resource tool.

For Doc Ligot, it is a “war between machines and humans,” with AI increasingly becoming better.

Thus, with the advent of these new technologies, he is stressing the need to put in place the necessary legislation to punish and prevent abuse in line with existing cybercrime and data privacy laws. On the other hand, Doc Ligot also laments, “we don’t have incentives to do AI.”

Doc Ligot points out that countries like our ASEAN neighbor Malaysia are training 30,000 engineers, while Middle East powerhouse Dubai is training a million prompt engineers. Prompt engineering is the process of structuring an instruction that can be interpreted and understood by a generative AI.

Funding for AI adoption and training, Doc Ligot noted, is hard to come by as the Department of Science and Technology’s (DOST) budget allocation is equivalent to less than 0.2 percent of GDP when the ideal should be at one percent.

Also problematic is any request for funding or grants, with a very long processing and lag time for the approval of such request. However, Cong Joey explained that there is funding available under grants and aids but admits that bureaucracy does get in the way.

Doc Ligot suggests therefore that the government should all the more adopt AI technology to improve government efficiency. But again, the stumbling block is the high cost of acquiring and maintaining AI technology through the so-called “cloud.” Cloud storage is a mode of computer data storage in which digital data is stored on servers in off-site locations.

On the aspect of legal questions about the “liability” of AI with regards to its response or action, Doc Ligot informed us that the liability question is still not clear, with only the European Union implementing some rules while the stance of the US and Japan is still on deniability. He cited a previous case involving a self-driving for hire vehicle that was involved in a car accident.

The legal case was hard to prosecute over who was responsible - the driver who was not driving, the for hire vehicle operator, the car manufacturer or the algorithm. Thus, the issue of liability still has to be pinpointed.

Our country, unfortunately, Doc Ligot said, is “obsessed” over the bad news on scams and “deep fakes,” but while true, he warned, could leave us once again well behind the world which is already preparing to embrace this new technology.

The Descartes Institute, he cited, has ranked the Philippines’ digital preparedness at 82 out of 124, with digital skills at a disappointing zero, suggesting problems with infrastructure, digital literacy and overall readiness to embrace AI.

Another study he mentioned, by Oxford Insights, downgraded the government’s AI readiness at “70 plus out of 80 countries,” primarily due to deficiencies in human capital, highlighting issues in education and training that could impede the country’s ability to leverage AI effectively.

But it isn’t all bad reviews, Doc Ligot noted, because other institutions like the World Bank ranks the Philippines at No. 4 in the world in terms of traffic to AI, LinkedIn has said 86 percent of our knowledge workers use AI, and Electronic Hub reported last year that the No. 1 per capita use of AI is by Philippine BPOs.

Cong Joey, however, interjected that the use of AI by BPOs may not necessarily reflect “indigenous” use by Filipinos.

Ligot acknowledged that there are fears that AI could eventually wipe out Philippine BPOs, but current data also shows mixed results with only eight percent of BPOs reporting some displacements, while 13 percent indicated an increase. Thus, the mixed results, he said, indicates that “adoption is not universal, it’s in pockets...and it’s because we are too obsessed with negativity.”

There was so much more discussed and should bediscussed, but those will have to be tackled in another column.