2024 marked a big year for democracy, with elections held in over 70 countries representing nearly half the world’s population. It was also dubbed the year of the “AI election” as experts warned of the havoc that AI-generated disinformation could wreak. In 2023, AI expert Oren Etzioni told Fortune, “I expect a tsunami of misinformation,” and journalist Maria Ressa said the world is facing a “tech-enabled Armageddon.”

Leading tech companies even created the Tech Accord to Combat Deceptive Use of AI in 2024 Elections, citing concerns about how the “intentional and undisclosed generation and distribution of Deceptive AI Election content can deceive the public in ways that jeopardize the integrity of electoral processes.” Rest of World reporting over the course of the year suggests those concerns may have been exaggerated — or misplaced.

At the beginning of 2024, Rest of World set out to create an AI Elections Tracker with a goal of understanding the different ways that AI tools were being used, with a focus on countries outside the West. Throughout the year, reporters from across the world gathered unique and noteworthy instances of the ways that AI — primarily generative AI — was being used around elections. By the end of the year, the tracker included 60 entries from 15 countries. The entries note the type of media created with AI, including text, image, audio, and video; the platform(s) the content was posted to and spread on; and the country in question. 

Rather than a comprehensive database, our AI Elections Tracker compiled a handpicked data set, which offered a glimpse into the myriad ways politicians and citizens used AI to create political content leading up to — and in the aftermath of — elections. Most of the posts we collected are concentrated in the first half of the year because that’s when unique use cases first emerged and when many major elections took place.

Toward the end of the year, experts noted that people’s worst fears about the negative effects of AI on elections hadn’t materialized. AI researchers wrote in the MIT Technology Review that earlier concerns that an onslaught of deepfakes and other AI-generated misinformation would mislead voters were overblown. Meta reported in December that it didn’t find much AI-generated misinformation on its platforms and it was able to either remove or label it quickly. 

Experts who spoke to Rest of World said that while AI didn’t have catastrophic effects that swayed elections, there are still reasons to be concerned as fake AI content did reach wide audiences and social media users aren’t always adept at distinguishing AI content from authentic content.

“When it comes to AI and elections [in 2024], I’d literally seen the phrase, ‘It’s the dog that didn’t bark,’ but I actually think that is a significant overstatement,” said Vivian Schiller, executive director of Aspen Digital, an Aspen Institute program focusing on technology.

Witness, a Brooklyn-based nonprofit that helps people use technology to document human rights violations, also tracked the use of deepfakes around elections last year. “What we saw was that there were clear faultlines emerging around the capacity of journalists, fact-checkers, and civil society to deal with, detect, and communicate around the deepfake cases they had,” said Sam Gregory, the organization’s executive director. He noted that AI detection tools didn’t always work effectively to provide conclusive results, and they didn’t have the capacity to deal with different accents or skin tones.

In India, there were many examples of people using generative AI to create memes in the lead-up to the country’s June general election. Nilesh Christopher, a former Rest of World reporter who tracked incidents in India, has been covering AI-generated election content since 2020. In 2024, he noticed that early instances of AI-generated content were memes, like a video of Prime Minister Narendra Modi transposed onto rapper Lil Yachty using Viggle AI.

“The playful use of AI or the engagement-oriented use of AI was higher compared to outright political uses or harms,” Christopher told Rest of World. “It became another tool to generate content on the cheap.”

Some of the most noteworthy AI use cases were for communication by politicians. A few weeks before Pakistan’s election in February, for example, the Supreme Court banned members of former Prime Minister Imran Khan’s Pakistan Tehreek-e-Insaf party from using its traditional symbol. All PTI candidates had to use new, individual symbols on ballots, creating confusion among voters. In response, the PTI created a chatbot to inform citizens who their local candidate was. After the election, Khan delivered a speech from prison with the help of AI as well. 

“One of the most powerful ways that particularly generative AI and image-making can do is it can turn metaphorical speech into imagery,” Gregory told Rest of World. “And I think that’s so obviously what is happening with so much of the political participation in a number of geographies.” In South Africa, for example, social media users used AI to create images of what the country would look like under Julius Malema’s leadership. In countries like Indonesia, Sri Lanka, and India, users and politicians resurrected political figures with AI for commentary or campaign endorsements.

Most of the incidents Rest of World tracked fell into two categories: memes and commentary by citizens, or political communication by candidates. But we also documented harmful uses of generative AI. In Mexico, for example, someone used generative AI to create a fake nude photo of candidate Mariana Rodríguez Cantú. In Taiwan, a Beijing-linked actor had disseminated AI-generated videos with disinformation about former President Tsai Ing-wen a week before the election.

Even content that seems less harmful at first can have negative consequences. AI-generated posts jumping from one platform to another, or reposted on the same platform, can lead to context collapse, where content meant for one group of people spreads beyond that bubble and gets misinterpreted. For example, one social media user used AI to create photos of Indian party leaders supporting their opponents’ policies. The original post on Instagram was posted with the caption “Politics in a parallel universe” and included a disclaimer that the images were AI-generated. But these images could be misleading if they were shared without this context. 

Shukri Hassan, who analyzed South African content for Rest of World’s AI tracker, saw a similar situation play out in her own life. While she was able to immediately recognize an AI-generated video of former President Joe Biden threatening sanctions on South Africa, some people in her personal orbit thought it was real. 

“For the most part, I don’t think the technology has yet advanced enough to mislead people convincingly to the point where they would change their votes,” said Hassan. “But especially with the older generation, it’s difficult for them to comprehend that someone can just load an old video onto a platform and create this kind of [fake] visual.”

A new report out of Indonesia, based on surveys conducted by the Luminate foundation and market research firm Ipsos, underscored concerns about the general public being unable to identify AI-generated content. (Luminate, which funds projects to support governance and citizen engagement issues globally, underwrote Rest of World’s coverage of the intersection of tech and democracy in 2024.) Of the 1,000 respondents, 63% said AI-generated content had the potential to impact their own personal political views. But of the 33% who said that AI-generated content had little or no impact on their personal views, “42% admit they are not very/at all confident in identifying AI-generated content,” the report said.

In a December report about global election content on its platforms, Meta noted that AI content relating to elections represented less than 1% of all fact-checked misinformation. A month later, the company announced it was shutting down its fact-checking programs and moving to X-style community notes. 

Schiller, from Aspen Digital, noted that Meta — which owns WhatsApp, Facebook, and Instagram — is the internet in many places around the world. Meta itself noted that its “existing policies and processes proved sufficient to reduce the risk around generative AI content” in the 2024 elections. So, social media platforms moving away from fact-checking and labelling content as AI-generated “is extremely troubling,” said Schiller. “[X] has become very unreliable. Meta has stopped its content moderation. These are a lot of the crucial tools that people use.”
The 2024 elections were a “dress rehearsal,” Schiller told Rest of World. “2024 was a year that was unique in terms of the volume of national elections around the world, unprecedented. But the elections keep going,” she said. “There’s another dozen or more national elections that are happening in 2025. And so the sophistication [of AI] is just going to get greater.”

India’s homegrown ride-hailing startups are forcing Uber to rethink its strategy.

In February, the Silicon Valley giant ditched its commission-based pricing for two- and three-wheeler segments, allowing drivers to pay as little as nine rupees (11 cents) for a daily subscription.

An Uber spokesperson said the company had recognized the market trend set by rivals and had adopted a similar strategy in the low-cost mobility segment to remain competitive. The spokesperson did not comment on whether Uber would replicate its commission-free policy for cabs. In the cab segment, drivers say Uber has slashed fares by 40% to retain its customer base, leaving them with little cash after completing rides.

This can lead to changes in the policy regulations and reforms in the Indian transport sector.

Mobility experts believe the shift to subscription models could alter ride-hailing economics globally. But as Uber cuts fares to hold on to its customers in India, they warn that deepening price wars could stifle the growth of upstarts and worsen drivers’ working conditions. Drivers, meanwhile, believe the innovative pricing models are temporary bait to lure them. 

“Any uptick in growth, particularly in offering different platforms with varying attributes, is an interesting development for a large market like India,” D. Dhanuraj, a mobility expert and founder of an Indian think tank, Centre for Public Policy Research, told Rest of World. “This can lead to changes in the policy regulations and reforms in the Indian transport sector. If that happens, there will be a lot to learn from the Indian market.”

Uber leads India’s cab-hailing segment with a 50% market share. In the popular scooter and three-wheeler markets, though, the company lags local rivals like Rapido, a unicorn that controls 31% of the market. Uber’s longstanding Indian rival Ola holds 26% in the same category.

Namma Yatri, a homegrown ride-hailing company founded in 2022 that pioneered commission-free pricing, operates on a subscription model: drivers pay a fixed daily fee — varying between 10 rupees (12 cents) and 30 rupees (36 cents), depending on the city — to use its app, and they get to keep the rest of their earnings. Namma Yatri has logged nearly 75 million rides since 2022.

Shan M.S., founder of Namma Yatri, told Rest of World that driver resentment toward commission-based pricing has been deeply entrenched — and palpable — for many years.

“When we spoke to drivers [in 2022], they demanded trust and transparency in payments, feeling deceived by commission models where algorithms set rates during peak hours,” Shan told Rest of World.

Rapido, another homegrown ride-hailing service, which primarily operates as a bike taxi aggregator, completes an average of 2.5 million rides daily and serves over 31 million active users.

Uber has clocked 3 billion rides in the last 10 years in India, an average of 800,000 trips per day.

The competition in the commission-free ride-hailing industry is only increasing.

In late March, India’s home minister, Amit Shah, introduced a new ride-hailing initiative called Sahkar Taxi in Parliament. Shah said the government-run service would be “just like Uber and Ola” in its reach and market size but would operate on a commission-free model.

The Sahkar Taxi initiative is part of a government effort to curb the dominance of large global tech platforms. It complements another public technology initiative — the Open Network for Digital Commerce (ONDC), launched in 2022. ONDC promotes a decentralized, open-source online commerce model that allows buyers and sellers to transact across different platforms without being tied to a specific app or website. 

Namma Yatri became the first ride-hailing platform to join ONDC in March 2023. A year later, Uber signed a pact with ONDC to explore potential integration, which hasn’t happened yet. 

“Uber has invested heavily in their proprietary technology, and they may not easily give up their edge by joining the ONDC platform,” Dhanuraj said. “ONDC is a decentralized model, and Uber is a centralized model. Uber is also facing many challenges in many states, and I don’t think Uber wants to get exposed to government oversight.” 

Kishan Verma, who leads an auto drivers’ union in New Delhi, told Rest of World that the way forward for the government is to set fare guidelines for ride-hailing apps. Though he’s not fully aware of ONDC, he’s advocating for a similar intervention.

“This is a welcome change from Uber — it was much needed — though we do not trust any of these aggregators, as once we get hooked, they will spike their rates,” Verma said.

Shaik Salauddin, national general secretary of the Indian Federation of App-Based Transport Workers, also sees zero-commission platforms as temporary bait.

“Some of these companies began with free subscriptions,” he said, “but once they attracted drivers, they started hiking fees.” In Hyderabad, Rapido now charges 29 rupees (35 cents) a day for three-wheeler taxi drivers, while Uber, still new to the model, is offering discounts and charging just nine rupees (11 cents), he said.

These shifts could dent Uber’s short-term revenue, Shreya Gadepalli, a mobility expert and former South Asia director at the New York–based Institute for Transportation and Development Policy, told Rest of World.

“But we’ll have to see how the subscription model plays out,” Gadepalli said. “The SaaS approach removes the uncertainty of fluctuating commission-based earnings.” 

Mohammad Aslam, an auto-rickshaw driver with Uber in New Delhi, said the change has brought relief. “The old system of paying a commission on every ride forced many of us to go offline and look for passengers on our own,” he said. “We don’t have to do that anymore.”

Aslam used to pay Uber 19 rupees (22 cents) per ride — nearly 30% of his daily income. Under the new subscription model, his flat fee is under 5% of what he earns in a day — a savings that, he says, covers his daily groceries.

Sachin Sachdev, an Uber Premier cab driver in Delhi, said Uber’s response to challengers like Rapido and Namma Yatri has been to slash fares across categories — from its Premier cars to the budget segment, UberGo. But the company still charges a 25% commission per ride for cabs, making the fare cuts harder to bear.

“Uber knows slashing prices is the only way to keep customers hooked,” he said. “It slows down Rapido and others. Drivers feel compelled to come back to Uber — that’s still where the ride demand is.”

As U.S. President Donald Trump sends mixed signals on the H-1B visa program, Amazon could find itself in a tight spot.

The $2 trillion tech giant has bagged the most H-1B visa approvals each year since 2020, according to data from United States Citizenship and Immigration Services. The USCIS data includes newly issued visas and renewals. The H-1B visa allows highly skilled foreign workers to live and work in the U.S. for up to six years, after which they can apply for a green card and be eligible to become naturalized citizens.

In 2024, Amazon Com Services obtained at least 9,265 H-1B visas. Many of its subsidiaries, such as Amazon Web Services and Amazon Development Center, also obtained visas in this category.

U.S. tech giants have become increasingly dependent on H-1B visas as they “have grown exponentially over these past few years, and they’ve added many more verticals … so they need more IT people for that,” Poorvi Chothani, founder of Mumbai-based global immigration law firm LawQuest, told Rest of World.

The Trump administration is split on the visa program, with key figures like the president’s homeland security adviser and deputy chief of staff, Stephen Miller, pushing for severe restrictions. Tesla CEO Elon Musk, as an informal adviser to Trump, has advocated for its preservation.

Trump has shifted his position on the visa in the past. In 2016, he vowed to “end forever the use of the H-1B as a cheap labor program.” In late 2024, however, he said, “I’ve been a believer in H-1B. … It’s a great program.”

On January 17, the Department of Homeland Security introduced a slew of measures to tighten the H-1B protocol — the most notable of which requires employers to submit more evidence for proof of control over foreign employees to mitigate alleged misuse of the program. The department also increased the non-refundable fee per entry to $215 from $10 during this year’s filing cycle.

USCIS has set the annual cap for new H-1B visas at 85,000, distributing them through a lottery system. More than two-thirds of the visas go to tech workers each year. Indian tech workers typically secure more than 70% of H-1B visas.

“The demand for these visas is still there and companies are just rolling with the changes. They are also filing additional paperwork and providing enhanced information in the H-1B petitions to mitigate the risk of denials,” Chothani said.

For most of the 2000s, Indian IT companies dominated the H-1B visa program, securing thousands of approvals. By 2024, the balance had shifted: Five U.S. tech giants — Amazon, Google, Meta, Microsoft, and Apple — were among the top 10 companies to get the H-1B.

Over the years, Democrats and Republicans have both proposed tightening the program to save American workers from being “ignored or unfairly disadvantaged.”

Tech professionals around the world are on the edge as President Donald Trump and his administration impose a series of radical immigration measures.

From attempting to deport foreign nationals who are permanent legal residents to barring visitors from entering the country for expressing views on U.S. foreign policy, immigrants have seen a tumultuous couple of months.

Last week, reports said the U.S. government is considering placing travel restrictions on citizens from 43 countries. The plan would completely ban travel from 11 nations, impose heavy curbs on 10 others, and give 22 countries a 60-day ultimatum to address perceived deficiencies or risk moving to one of the other categories.

While workers from the targeted nations make up a relatively small percentage of the tech workforce, industry professionals outside the purview of the measures, too, have been increasingly concerned about being impacted by the shifting visa policies and anti-immigration rhetoric.

They are particularly worried about increased scrutiny of the H-1B visa, which is sought after by tech companies.

Immigration attorneys are now advising clients on temporary U.S. visas to avoid leaving the country, warning they may encounter difficulties when attempting to return.

“There has been a surge in audits and enforcement actions against employers hiring foreign workers, creating job insecurity for legal immigrant employees,” Poorvi Chothani, managing partner at immigration law firm LawQuest, told Rest of World. U.S. Immigration and Customs Enforcement officers have been increasingly visiting employees’ homes to “check if they are legally working from home,” she said in a statement. Chothani’s firm represents hundreds of individual and corporate clients.

Rest of World spoke with tech workers across several countries to record their concerns. Several of them asked to remain anonymous, fearing backlash from the U.S. visa authorities.

A manager at a Chinese tech company, China

I have a 10-year visa, so I do not think this will affect me. But I want to get my husband a visa, too, and I am worried that his application will be rejected. I heard it is hard to get even a tourist visa right now — people seem to be getting rejected for no reason. 

I own property in the U.S. and if I am there to travel and visit friends, I would surely like to go with my husband. But given the situation, we wouldn’t dare to apply right now because if there is a record of him getting rejected once, it would be harder to get a visa in the future.

I realized that the immigration system would always be a noose around my neck.

Atal Aggarwal, founder of immigration-focused software development firm OpenSphere, India

I lived in the U.S. on H-1B for seven years. I wanted to build my own company and realized that the immigration system would always be a noose around my neck. I wanted the freedom to do what I wanted but it felt like it was a system where I kept pumping cash to get the right legal status. It just wasn’t working out for me.

Around November last year, I thought I should take control of my life and work on achieving my dreams instead of waiting for the U.S. government’s help.

I returned to India, and since then, I have seen a lot of community news about the struggles Indians are facing with the immigration system without a proper way forward. People are realizing it’s not worth the stress and that it’s not going to get better for the next four years until Trump is out of power. So they are moving ahead — either returning home or shifting to other countries where they can have a better life without the worries about immigration systems.

A Nigerian entrepreneur who lives with his family in Austin, Texas

One of my friends who lives in the U.S. got fired last weekend. Now, he has just 30 days to find another job or leave the country. He relocated from Nigeria to the U.S. with his wife and children last year. They sold everything they owned back in Nigeria before moving.

There are several immigration changes already affecting business or will make it harder to do business in the U.S. in the future. For example, the U.S. “Gold Card,” which everyone is talking about, is hard for any entrepreneur to get because of global taxation norms.

Even though I am on a green card right now and the recent visa changes have not affected me, the truth is that I need to rethink my future. I will definitely relocate out of the U.S. and I’m planning to start looking for another country.

An Indian tech worker employed at a big tech firm in California

I hold a green card but my family back in India is worried that if I travel to meet them, I might face difficulty reentering the U.S. My family is even worried about anything I post on my social media.

I had a son six months ago and we wanted to take him to meet his grandparents in India. Typically, our parents would be excited about it but this time, given the fears around visas and layoffs in Silicon Valley, they asked if they should come to the U.S. because “what if you guys come here and they never let you back in?”

I am finally traveling back later this week, and my dad is constantly calling me to ask me not to take anything lightly, to stay safe, to carry my old green card and visa paperwork, and not leave anything to chance. It’s not advice I am used to hearing from my father. I am used to him saying, “Come on. Life will be good.”

I don’t even know if I want to live in the U.S. for the rest of my life. My personal bias towards the American Dream is at risk.

An Indian tech worker at a tech firm in the U.S.

I am on an H-1B and I’ve reached a stage where I feel like if this doesn’t work out, I’ll just go back to India. My sister’s visa to visit me was rejected last week, which is something I had never heard of happening before.

I don’t know if it’s about the last two months as much as it is about the past 10 years. Being stuck in the H-1B loop for an Indian is always bad. What Trump can do is make certain things harder. But he is not the one who introduced the deathly 18- or 20-year waits for green cards. That has always been the reality of H-1B workers. But how much harder can they make it for someone stuck in the 18-year timeline for a green card?

My personal bias towards the American Dream is at risk.

An Indian tech consultant with a large consulting firm, U.S.

It is a strange situation. I know of people making hundreds of thousands of dollars but they know that they will have to move to Canada next June because their visa won’t be extended. In any other situation, if you made that kind of money, it would be wild to imagine that you can’t even stay where you want to stay.

It restricts how you make your life decisions or travel or see your family. People can’t get married or fly to attend a funeral.

The correct term to describe my status in the U.S. is “a non-resident alien.” I am not an illegal alien, but I am an alien nonetheless. I am not a citizen so I don’t have a voice in how things function. It may seem unfair but you voluntarily sign up for it and most of us don’t really think it through in the hope of a better life.