Regulating A.I.

By James Jimenez

With the 2025 National and Local Elections just around the corner, the recent quantum advances in the sophistication of AI products – particularly deepfakes – has occasioned more than a little handwringing in the halls of Congress. All to the good, of course, as legislators – led by Rep. Luis Raymund Villafuerte – have introduced House Bill 10567, which aims to regulate AI amid the threat posed by the proliferation of deepfakes or digitally altered images and audio or video recordings that misrepresent victims.

To my mind the bill, filed on August 12, 2024, reflects an appropriately proactive approach to safeguarding the integrity of the electoral system in the face of the blinding pace of technological advancements.

The Need for Regulation

Hardly anyone can question the need for regulation. AI-driven tools are being used to manipulate information, leading to the creation of life-like deepfakes, which are then used to sway public opinion. If you think this poses a significant threat to free and fair elections, you would not be wrong.

In the U.S. deepfakes are now being deployed with some regularity – and in some cases, by high profile public figures like Elon Musk. Musk claims he did it for the hilarity of it all, but the inescapable effect of releasing video featuring the AI-doctored voice of the Democratic nominee disparaging her self as a DEI hire, is to unduly sway the minds of voters, mere months before the Presidential elections.

And even if no AI has actually been used, the mere possibility of its utilization can be capitalized on as an attack line in the campaign. Very recently, for example, the size – nay, the very existence – of the crowds in VP Kamala Harris’ campaign stop in Michigan was questioned by the Republican nominee himself, former President Donald Trump, who claimed that crowds had been “A.I.’d” into the pictures of the event.

We have not seen similar instances in the Philippines, but that is probably only because we haven’t really gotten to election season yet. However, that will probably change sooner rather than later as the filing of Certificates of Candidacy for the 2025 polls is slated to begin on the first day of October.

With social media already playing an outsized role in our politics, potentially even to the point of affecting election outcomes, the unregulated use – and misuse – of AI could very easily exacerbate issues such as misinformation, voter manipulation, and the erosion of public trust in the electoral process.

Implications and Challenges

If passed, this bill will give the Philippines the distinction of being one of the first countries in the world to implement comprehensive regulations on AI in elections. The new law might even serve as a model for other nations grappling with similar concerns about the impact of AI on their democratic processes.

Getting the law passed and carrying out its provisions are two different things however. One of the major concerns about this proposed new law is the enforcement of its proposed regulations. Given the decentralization and anonymity that is possible with AI, I simply cannot conceive of effective policing without extensive training being given to law enforcement.  It would not be an exaggeration to say that ensuring compliance from both local and international actors involved in election-related activities will be a ludicrously complex task.

Nevertheless, the filing of this bill truly does represent a crucial first step towards the responsible use of AI in the Philippines, particularly in safeguarding the democratic process. As with the abrupt and wholesale transition to an automated election regime in 2010, the Philippines will once again have the eyes of the world upon it, standing as a pioneer at the intersection of AI and electoral integrity.