The paramount danger posed by unchecked artificial intelligence (AI) demands a global collective effort to control its evolution, declares Geoffrey Hinton, the former vice president of engineering at Google, also fondly christened the “godfather of AI”.
In a recent interview with EL PAÍS via videoconference, Hinton underscored the urgency to grasp the reins of this burgeoning technology before it outsmarts us.
Related: Scathing RNC Ad Uses AI to Show What Disastrous Biden 2nd Term Would Look Like – Video
Despite his recent resignation from Google, Hinton continues to sound the alarm about the potential risks of AI, urging humanity to recognize the necessity of its regulation.
He criticizes the call for a half-year moratorium on the training of AI systems more potent than OpenAI’s GPT-4 as “completely naive”. Instead, he insists that the brightest minds should focus on understanding and mitigating the risks associated with the technology.
“AI is an extraordinary technology, spearheading breakthroughs in diverse fields such as medicine, material science, and natural disaster predictions,” Hinton acknowledged. “Yet, we must comprehend and control it, ensuring we prevent any negative consequences before they materialize.”
He further recommends that governments enforce the flagging of falsified images created through AI, emphasizing the urgent need for transparency in the digital world.
In the face of the monumental task of developing AI technology while ensuring its safety, Hinton confesses that the current efforts are falling short.
He questions the feasibility of this dual mission in a capitalist system, underscoring the inextricable intertwining of economic and technological concerns.
When it comes to the shared apprehensions within his circle, Hinton confirms that many highly intelligent individuals are “seriously concerned”. “We’ve entered uncharted territory, building machines stronger than ourselves while maintaining control—for now. What happens when we build machines smarter than us?” he challenged, expressing his fears about our lack of experience in handling such prospects.
Hinton points out multiple threats posed by AI, including job displacement and the proliferation of fake news. With AI models like ChatGPT processing vast amounts of data beyond human capacity, he suspects AI might be surpassing human efficiency in various tasks. “That’s what scares me,” he confessed.
Speculating on the timeline for AI to outstrip human intelligence, Hinton provides a rough estimate of five to twenty years, admitting uncertainty about the prediction.
The interview brought forth a significant question—would AI eventually have its own goals? Hinton believes this is a critical concern, the “alignment problem”, referring to the challenge of aligning AI’s goals with human interests. He warned about the dangers of robot soldiers and the possibility of AI self-generating targets, thus becoming more powerful to achieve its objectives.
Despite lauding Google’s responsible behavior, Hinton emphasizes the reality of businesses operating within a “competitive system”. He questions the capacity of the United States’ political system to effectively regulate AI, given its struggles with issues such as firearm control.
Related: Biden Signs New Executive Order that Experts Claims Could Promote “Woke” Artificial Intelligence
“I’m generally optimistic, but the U.S. political system’s inability to make a decision as simple as not giving assault rifles to teenagers doesn’t inspire confidence in handling a much more complex problem like this one,” he expressed, highlighting the need for political reform to address this technological challenge.
Hinton emphasized the urgency to prepare for the looming challenge of AI. “There’s a chance that we have no way to avoid a bad ending … but it’s also clear that we have the opportunity to prepare for this challenge. We need a lot of creative and intelligent people. If there’s any way to keep AI in check, we need to figure it out before it gets too smart,” he urged.