When Will AI Take Over the World? We Asked ChatGPT

Artificial intelligence (AI) is rapidly transforming political campaigns, raising concerns about its potential to undermine elections, particularly when it is used to generate deepfakes and misinformation.

The upcoming presidential election faces heightened scrutiny over these risks as AI tools become more sophisticated and accessible.

One high-profile example is a deepfake of President Joe Biden, which falsely depicts him telling people not to vote. This type of content, if unchecked, could damage public trust in democratic process.

Manipulated videos or audio clips can deceive voters by fabricating realistic but false statements by political figures.

In response, fact-checking organizations like Full Fact have started using AI tools to monitor claims in real-time, analyzing speeches and manifestos to prevent the spread of misinformation. However, the challenges remain significant, especially when deepfakes spread rapidly through social media, often before they can be debunked by eagle-eyed fact-checkers who can spot a fake video.

Mustafa Suleyman, cofounder of DeepMind, emphasizes these dangers in his book The Coming Wave, arguing that unchecked AI could significantly disrupt democratic processes. He advocates for stricter regulation and oversight to prevent AI technologies from being weaponized to mislead or manipulate the electorate. Suleyman’s concerns align with broader academic opinion, which warns of AI’s potential to destabilize elections globally if not properly controlled.

Polling also indicates public awareness of AI’s potential harm Almost half Americans fear AI could attack humanity. Conducted by YouGov, the survey revealed how Americans felt computer and human intelligence stacked up, and highlighted concerns over a possibility of conflict between man and machine.

Newsweek asked the generative AI tool, ChatGPT: “When Will AI Take Over the World?”

ChatGPT responded: “The question of whether and when AI might “take over the world” is a complex one that depends on how we define “take over” and what developments in AI technology might lead to such an outcome. It’s a topic often discussed in science fiction, but the reality is much more nuanced.

“Today, AI is powerful but highly specialized. It excels in tasks like image recognition, natural language processing, and data analysis, but it is not close to matching the full spectrum of human intelligence. We refer to this level of AI as narrow AI or weak AI, meaning it is good at specific tasks but lacks general cognitive abilities.

“The idea of AI “taking over” usually refers to artificial general intelligence (AGI) or superintelligence—AI systems that can perform any intellectual task a human can and possibly surpass human intelligence. We are nowhere near developing AGI. Experts are divided on when, or even if, AGI will be achieved. Some believe it could happen within a few decades, while others argue it may never happen.

“Optimistic View: AGI could lead to massive advancements in technology, medicine, and quality of life, but it would need to be carefully managed.

“Cautionary View: Without proper safeguards, AGI could become uncontrollable, leading to unintended consequences, especially if it surpasses human intelligence without being aligned with human values.

Risks and Challenges

“There are several potential risks, but they are far from certain.

“One near-term risk is AI being used in military applications, like autonomous drones, which could cause serious global instability.

“Job Automation: AI could disrupt the workforce, automating many jobs, which might lead to societal upheaval if economic and social systems don’t adapt.

“Ethical and Control Issues: The biggest concern for AGI is alignment—how to ensure that AI’s goals are aligned with human values. If AGI is not well aligned, it could theoretically pursue goals that are harmful to humanity, either deliberately or as a side effect of its programming.

Control and Safety

“Organizations like OpenAI, DeepMind, and others are actively researching how to build safe and controllable AI systems. There are various approaches to ensuring AI remains beneficial, including:

Value alignment: ensuring AI systems understand and follow human ethical principles.

Regulations and Governance: Developing international agreements to control the use and development of AI, especially in dangerous applications.

Timeline Speculation

“Short-term (Next 5-10 years): AI will continue to make significant strides in specific fields such as health care, transportation, and education. However, we will still be working with narrow AI systems, and they are unlikely to “take over the world” in any sense.

“Midterm (20-50 years): Some researchers speculate that AGI could emerge, but this is speculative and highly uncertain. If it does, it will come with major ethical and safety considerations.

“Long-term (50+ years): Predicting AI development this far out is extremely difficult. If AGI and superintelligence are developed, careful management will be crucial to avoid harmful outcomes.

Conclusion

“AI is unlikely to “take over the world” in the near future. While AGI, or superintelligent AI, could pose risks if developed, we are still far from reaching that stage.

“Whether AI becomes a force for good or danger will depend on how humanity shapes its development through ethical guidelines, governance, and safety measures.”

Do you have a story we should be covering? Do you have any questions about this article? Contact LiveNews@newsweek.com