AI programs can aid criminals but that is not the real danger

It was on November 30, 2022, that the California-based OpenAI company released its artificial intelligence-powered bot, ChatGPT. Within a span of nine months, Kerala itself recorded the first AI-linked scam, with a person in Kozhikode falling victim to an AI-generated video phone call.

The Kerala police have now extended their probe into two states, and what will come of it is not clear. However, what is evident is that criminal elements have quickly understood the power of AI, which can create high-quality images, videos, and text.

Manipulated images and fake narratives have long circulated through social media, pushing some societies to dangerous levels of polarization. But AI introduces two scary elements: first, there is no need for any technical expertise, as AI can create deep fakes with just a typed command, eliminating the skills needed to use a program like Photoshop. Second, it can be done in seconds and on an industrial scale.

We should have realised this when a viral clip surfaced, showing Mohanlal and Mammootty portraying characters from Francis Ford Coppola's iconic movie, 'The Godfather.' At first glance, it looked like cool and harmless fun. Yet, the message posted by the creator of the 'Godfather' clip, YouTuber Tom Antony, was a clear warning. In a video posted soon after the clip went viral, he revealed that some people who contacted him were eager to know how to make similar videos.

Now, consider the 'Godfather' clip from a different perspective. It was a video created by merging the faces of the actors into a scene from a Hollywood movie, without any of the actors knowing or giving consent. Do you see the danger now? Anyone's picture or voice can be taken from the internet and used to create a highly convincing replica of their voice or their face morphed into a video.

The rapid advancement of AI capabilities in recent months has enabled the generation of realistic videos using just a few photographs, making it seem as though a person is mouthing words they never actually said. While this technology has been evolving for some time, it has now reached a level of sophistication that even experts find challenging to distinguish between real and fake.

As expected, individuals with malicious intent have discovered the potential of using such clips in messaging apps to deceive people. We can anticipate more scams of this nature to surface. Reports have already emerged from the United States and China about scammers using the voices of family members to dupe elderly parents into sending money. The FBI issued a public warning about these scams in June.

According to the FBI alert, victims of these scams include not only adults but also minor children and non-consenting individuals whose content on social media is altered into explicit material. The manipulated images and videos are then spread widely on social media platforms or even on pornographic websites.

As mobile phones and messaging apps have become integral parts of our daily lives, the opportunities for criminal elements to exploit such technology are plentiful.

Superintendent of Police with the Vigilance and Anti-Corruption Bureau, ES Bijumon, warns that people must exercise extreme caution when receiving unusual calls from family members. It's crucial to verify the authenticity of such calls by reaching out and confirming with the family member before taking any action.

Bijumon emphasises that any request for money transfer to unfamiliar accounts should be considered a red flag, and individuals should immediately get in touch with the police before sending money to strangers in distant locations. He acknowledges that even educated individuals may fall into such traps when panic sets in.

According to Bijumon, criminal gangs are becoming increasingly cunning in devising convincing scams, as demonstrated in the Kozhikode case. He reveals that some syndicates keep track of people who are traveling and exploit this information to make calls to their relatives during their absence. Taking advantage of the unavailability of travellers, these criminals launch their fraudulent schemes.

Bijumon also highlighted that most travellers tend to use Wi-Fi networks while at airports, providing an opening for hackers to steal sensitive information as these networks are often not secure.

"It is not just private individuals who are targeted. Business firms are also at risk, as employees receiving phone or video calls purportedly from their superiors may unwittingly disclose trade secrets that could cost the company millions," warns Bijumon. Therefore, it is essential to treat anyone asking unusual sensitive questions with suspicion, not just those seeking money.

This security concern also extends to the military and security forces. During an interaction at the Technopark earlier this year, Brigadier Lalit Sharma, who heads the Pangode military station in Thiruvananthapuram expressed that the Indian army is actively focused on finding centralized solutions to combat this issue.

"We have also sensitised our people to be extremely wary of such attempts being made on their personal calls. We are actively conducting a major drive to educate our troops, families, and children, as they might have access to location and other sensitive information," said Brigadier Sharma.

Another area where deep fakes are expected to cause havoc is politics. With over 1 billion people poised to vote in 2024, including India and the US, experts fear a tsunami of deep fakes.

Microsoft founder Bill Gates issued one of the clearest warnings on this issue: "Deepfakes and misinformation generated by AI could undermine elections and democracy," Gates recently wrote. "On a larger scale, AI-generated deepfakes could be used to manipulate election outcomes. Although it doesn't require sophisticated technology to cast doubt on the legitimate winner of an election, AI will make it easier," he said.

The possibilities are indeed concerning. With IT cells and cyber troll armies already adept at silencing and maligning opponents, the integration of AI-powered programs would provide them with a powerful tool. In just a matter of minutes, they could create videos depicting their political adversaries in compromising situations or making harmful statements and propagate them through social media platforms.

On the other hand, the potential for misuse is even more alarming. What if someone is captured on video engaging in inappropriate behavior or making controversial statements, but they claim it was fabricated using AI as their defense? This may sound ridiculous, but it has already become a reality. In the case of the January 6 attack on Capitol Hill in Washington, two individuals charged in connection with the incident tried to assert that the videos showing them at the scene were AI manipulated. However, the court rightfully rejected their claim.

Video clips have played a crucial role in exposing the grim reality of events like the situation in Manipur state, where authorities sought to suppress the truth by suspending the internet and creating an alternate narrative.

Now, imagine if they had indeed claimed that the video was fake and AI-generated. Given the lengths to which political parties in power go, this scenario is highly plausible.

Hany Farid, a digital forensics expert and professor at the University of California, has issued a warning about the potential impact of AI, and it should make all of us concerned about what the future holds. He expressed deep worry about a world in which evidence of "police violence, human rights violations, or a politician saying something inappropriate or illegal" could be easily dismissed because of the possibility of digital manipulation. "Suddenly there's no more reality," he warned.

This concern is already evident, even when receiving a phone call from a family member or a friend.