Is AI a friend, an enemy, or a partner?

Bicycles revolutionized how we travel, and now AI is doing the same with how we communicate and connect. AI is everywhere, silently operating within familiar applications we use every day.

From movie recommendations on online platforms and navigation apps to help you avoid traffic jams, to tools assisting doctors in diagnosing illnesses, AI has been changing the way we work, live, and interact. Its influence is ever-widening, opening up unprecedented possibilities but also posing numerous challenges.

The AI ​​Index 2025 report, published by the Artificial Intelligence Institute at Stanford University (USA), provides a comprehensive overview of the rapid development and far-reaching impact of AI globally.

This report synthesizes data from hundreds of sources, analyzing prominent trends in research, applications, investments, and policies related to AI, thereby painting a comprehensive picture of how this technology is reshaping many aspects of life, the economy, and society.

AI is everywhere and getting smarter.

We need to be acutely aware of the rapid pace of AI development and its ever-increasing penetration. In just one year, AI has advanced dramatically, from struggling with complex mathematical problems and exercises to being able to solve coding and scientific problems at the postgraduate level.

AI performance on skill tests such as the SWE-bench, a benchmark designed to test the ability of AI models to solve real-world software errors, has made incredible strides.

Accordingly, in 2023, AI could only solve 4% of these errors. But by 2024, it had solved nearly 72%, an extremely large leap in technical capability.

And these advancements are no longer confined to the laboratory. AI is gradually becoming an integral part of daily life. In 2023, the U.S. Food and Drug Administration (FDA) approved 223 AI-powered medical devices, compared to just 6 in 2015.

In many major cities around the world, people can now easily hail self-driving taxis from companies like Waymo in the US or Baidu in China.

Global investment in AI has also reached unprecedented levels, totaling $252.3 billion in 2024, a figure nearly equivalent to the entire GDP of Portugal in 2023 (approximately $289 billion).

To put it into perspective, this investment is 10 times larger than NASA's Space Center budget for fiscal year 2023 ($25.4 billion), and far exceeds the total cost of building the International Space Station over several decades.

Businesses are also jumping on the bandwagon, with 78% of companies in 2024 having adopted AI, compared to 55% just a year earlier. The U.S. still holds the lead in developing the most advanced AI models, but China is rapidly closing the gap, its leading models now approaching the quality of their American counterparts in key tests.

For users, the good news is that AI is becoming increasingly cheaper and more accessible. In just two years, the cost of many AI tasks has decreased by more than 280 times. Governments have also begun to invest heavily and draft legal frameworks to regulate the AI ​​boom.

The positive aspects of AI: supporting humans and society.

While the prospects of AI often attract a lot of attention, the reality is that AI has already been delivering tangible benefits to humanity and society, especially when combined with human skills and directed towards the common good.

One of the biggest advantages of AI is its ability to improve labor productivity. Numerous studies show that AI tools help employees accomplish more tasks, especially when assisting those with less experience, thereby narrowing the skills gap in the workplace.

Imagine AI as an "intelligent assistant" that handles repetitive tasks, allowing humans to focus more on creative work and complex thinking.

Beyond individual benefits, AI also promises to address major societal challenges. "AI for society" initiatives are leveraging this technology to serve the common good, from monitoring climate change and managing water resources efficiently to expanding access to healthcare in remote areas and personalizing learning experiences for students.

In science and medicine, AI is also making leaps and bounds. Tools like AlphaFold3 help scientists better understand protein structures, accelerating the development of new drugs at an unprecedented pace, with many achievements reaching Nobel Prize-winning levels.

AI has even surpassed human doctors in some specialized tasks, such as detecting cancer through medical images. In addition, AI tools are also helping doctors reduce administrative burdens and alleviate paper-intensive burnout.

However, public views on AI remain highly polarized. According to the AI ​​Index 2025 report, more than half (55%) of people globally believe that current AI offers more benefits than drawbacks, indicating cautious optimism.

But when considering the long-term outlook, especially in areas like the U.S., the picture becomes less rosy. A recent Pew Center survey in April 2025 found that only 17% of American adults believe AI will have a positive impact on the country within the next 20 years, while more than 50% expressed more concern than enthusiasm.

This difference shows that perspectives on AI can vary depending on geographic location and whether people are evaluating current achievements or long-term future impacts.

Humans must accept and confront the difficulties and pitfalls created by AI.

To appreciate the benefits that AI brings, we must also accept its downsides and risks.

Building trust in AI is a major challenge, especially as many people are concerned about how companies use personal data and the risks associated with bias in AI systems.

Even AIs designed to be neutral can inadvertently pick up and reproduce racial or gender biases from their training data.

Misinformation is a significant concern. AI can easily create deepfakes (fake images, videos, or audio), thereby creating opportunities to spread misinformation, especially in elections or for the purpose of personal harassment.

We are also seeing numerous reports of AI flaws, such as facial recognition leading to false accusations or chatbots offering harmful advice.

Furthermore, data collection and usage have also become a problem, as many websites are starting to block access, reducing the effectiveness of AI in the future.

Despite significant progress, AI still has many limitations. It still struggles to solve complex reasoning problems that require common sense or long-term planning abilities.

Training powerful AI models requires enormous amounts of energy, contributing to increased carbon emissions. While AI has already assisted many workers, concerns about job automation cannot be ignored. Embracing AI means facing these complex issues honestly and frankly.

Ensuring the future of AI is trustworthy and consistent with human rights.

Ultimately, we need to answer the question of who will be responsible for guiding the development and use of AI. While efforts to develop AI responsibly are increasing, we still have a lot of work to do.

Although progress has been made in testing AI for safety and fairness, this is still not a universal standard for all major AI developers.

Governments are becoming increasingly proactive, investing in AI technology and enacting regulations. Countries are beginning to cooperate more on AI safety, establishing international groups and agreements focused on making AI trustworthy and human rights-compliant.

However, companies often discuss responsible AI more than they implement truly robust safeguards. You might recall the AI ​​development moratorium that over 1,000 tech leaders, researchers, and other individuals signed in March 2023, but very few actually complied.

We need more transparency in how AI systems are built and trained. Each of us, from developers and businesses to governments and citizens, has a role to play in this.

Education is key. We need to ensure everyone understands AI, but the current awareness gap remains very large.

For example, in the US, although 81% of computer science teachers believe that AI should be taught in schools, less than half (only 46% of high school computer science teachers) feel ready to teach about AI.

Bridging this gap is crucial to building a future where everyone can navigate and contribute to shaping the world of AI. And in this context, we also need to remember that billions of people around the world have never had access to ChatGPT, not because of a lack of interest, but because they struggle with basic needs like clean water, nutritious food, or internet connectivity.

In summary, the AI ​​Index 2025 report marks a solid starting point in understanding the ever-changing landscape of AI technology. By maintaining awareness, appreciating positive values, confronting challenges, and recognizing our responsibility to ourselves and our communities, we can contribute to building a fair and inclusive AI future for all.