Worry about the AI catastrophe

Prominent figures in the field of artificial intelligence are concerned about the potential dangers posed by superintelligent AI, even predicting an apocalypse.

” There is both excitement and fear. The doomsday scenario of AI is that it may resemble the storyline depicted in the Terminator movie ”

Stated Ivana Bartoletti, a data protection and privacy expert, as well as the founder of the Women Leading in AI Network organization, in her assessment on The Guardian.

Since last year, Artificial Intelligence generative models such as OpenAI’s ChatGPT, Dall-E, Google’s Bard, and Midjourney have become widespread. However, these tools can produce biased outcomes, such as prioritizing men over women, inaccurately recognizing faces of people with darker skin tones, or assuming higher fraud risks for immigrant families. These issues can potentially push many individuals into dire situations.

“These are societal issues that we have recognized, demonstrating the need to achieve consensus on the right direction for AI, using this technology responsibly, and imposing constraints on them,” wrote Ms. Bartoletti.

Artificial Intelligence pioneer Geoffrey Hinton made headlines in early May as he announced his departure from Google. Hinton, the recipient of the esteemed 2018 Turing Award, renowned as the “Nobel Prize of Computing,” took to Twitter on May 1st to share his reason for leaving: “I’m leaving Google so that I can criticize Google. In fact, I’m leaving so that I can speak out about the dangers of AI without having to consider how this will impact the company.” His decision to distance himself from corporate affiliations allows Hinton to openly address the potential risks associated with artificial intelligence, generating significant interest within the tech community and beyond.

According to Hinton, the immediate threat posed by AI is the proliferation of fake images, videos, and texts on the Internet, as well as the profound changes it will bring to the job market in the near future. Ultimately, he is concerned that AI could become a threat to human civilization when it gains the ability to process massive amounts of data.

“When they can start writing code and running their own programs, assassin robots will come out in the real world. AI could be smarter than humans. Many people are starting to believe this. I was wrong to think that it would take another 30-50 years for AI to make this progress. But now, everything is changing too fast,” Hinton told The New York Times.

In mid-April, Google CEO Sundar Pichai also stated that AI keeps him up at night because it has the potential to be more dangerous than anything humanity has ever seen, and society is not ready for the rapid development of AI. “It could be a catastrophe if deployed wrongly. One day, AI will have capabilities that surpass human imagination, and we cannot fully fathom the worst things that could happen,” he said in an interview with CBS.

Sam Altman, the co-founder and CEO of OpenAI, shares a parallel sentiment regarding the potential impact of artificial intelligence. While recognizing the vast advantages AI can bring, Altman is vocal about the accompanying apprehensions it engenders. He emphasizes the risks associated with misinformation dissemination, potential economic disruptions, and the potential for AI to surpass human preparedness in certain domains. Altman’s concerns about AI are consistently voiced, and he openly expresses his astonishment at the widespread popularity of ChatGPT, acknowledging the need for vigilance and careful consideration as this technology continues to advance.

At the World Economic Forum event held in Geneva on May 3rd, Microsoft expert Michael Schwarz echoed the concerns surrounding the potential negative consequences of artificial intelligence. He emphasized that if misused, AI has the capacity to inflict harm, citing examples such as spamming, fraudulent activities, and even interference with electoral systems. Schwarz expressed his belief that malicious actors will exploit AI technology, underscoring the potential for significant harm to ensue. In light of this, he posed a fundamental question: How can we effectively regulate AI to harness its benefits and mitigate the risks? This pressing inquiry highlights the need for proactive measures to ensure responsible development and deployment of AI systems.

Experts are also looking towards a higher-level model called Artificial General Intelligence (AGI). According to a survey conducted by Stanford University last month, 56% of computer scientists and AI researchers believe that synthetic AI will transition to AGI in the near future.

Fortune states that AGI is considered much more complex than current Artificial Intelligence models due to its ability to possess self-awareness of what it says and does. In theory, this technology could lead to human concerns in the future.

According to the survey results, approximately 58% of AI experts view AGI as a “major concern,” and 36% believe that it could lead to a “nuclear-level catastrophe.” Some experts suggest that AGI may represent a “technological singularity” – a hypothetical point in the future when machines surpass human capabilities in a way that cannot be reversed and could pose a threat to civilization.

Demis Hassabis, the CEO of DeepMind, Google’s Artificial Intelligence company, expressed his astonishment at the progress of AI and AGI models in recent years. In an interview with Fortune, he stated, “I see no reason why the progress should slow down. We only have a few years, or at most a decade, to prepare.”

At the end of March, over 1,000 influential figures in the technology field, including billionaire Elon Musk, Apple co-founder Steve Wozniak, and renowned AI researcher Yoshua Bengio, signed a letter calling for global companies and organizations to halt the race for superintelligent AI for six months to collectively establish a set of common rules for AI.

According to Bartoletti, amidst the excitement and fear, the focus should be on finding solutions to control AI rather than halting its development. AI holds tremendous potential in various fields, such as analyzing cancerous tumors in CT scans. Last year, DeepMind’s AI successfully deciphered the structure of most known proteins, solving one of the biggest challenges in the field of biology in nearly 50 years.

Bartoletti emphasizes the need to approach AI with a balanced perspective and not to be blindly optimistic or excessively fearful. She believes that it is important to recognize the positive aspects of AI and to establish global rules if we want to harness its transformative power in a beneficial way.

Many governments around the world have started to pay attention to Artificial Intelligence regulation, including the United States, the European Union, India, and China. In a meeting with technology companies on May 2nd, President Joe Biden emphasized the need for these businesses to ensure the safety of AI products. On April 30th, the United Nations also announced the Global Digital Agreement, aimed at safeguarding human rights in the digital age.

These initiatives demonstrate a growing recognition of the importance of responsible Artificial Intelligence development and the need to establish frameworks and regulations to mitigate potential risks and ensure the ethical and safe deployment of AI technologies.