'We Think This Technology Is Too Important Not to Regulate': Google SVP James Manyika Warns AI Capability Must Not Outrun Safety
Google VP James Manyika said AI is too important not to regulate, arguing that oversight should involve industry, government and society. In a Bloomberg interview, he warned existential risks must be taken seriously and said AI capabilities should not outrun risk management. On jobs, Manyika rejected mass-elimination forecasts, saying most roles will change rather than disappear.
James Manyika, Google's Senior Vice President of Research, Technology & Society, feels that AI technology is too important not to regulate. He frames AI oversight as a responsibility shared across industry, government, and society rather than something left to companies alone.
In an interview with Bloomberg, Manyika didn't stop at calling for guardrails. He acknowledged the risks head-on, and said that "the dangers have always been real" and that AI's potential should be weighed alongside genuine concern rather than dismissed as hype. He went further on existential risk specifically, saying, "Any time there's any risk of existential crises occurring, you have to take that seriously." Underlying all of it, he said, is one core principle, "Capability development should not outrun our ability to manage risks and complexities."
ALSO READ
On jobs, Manyika pushed back against apocalyptic forecasts that have circulated since AI's current boom began. "Some of those predictions were made two years ago, that in two years, 50 percent of jobs would be wiped out. Well, two years is up. Let's take a look," he said. Rather than mass elimination, he described three parallel outcomes - some jobs shrinking, others growing, and most significantly, a large share simply changing in nature, since full automation of complex roles remains rare.
Manyika also offered a geographic contrast rarely highlighted in Western AI coverage, telling Semafor that the Global South is 'a little bit more optimistic' about AI than the US and Europe, pointing to real deployments including tuberculosis diagnostics tested in Zambia and India, Google Translate now covering 300 languages, and a Planetary Prediction Engine that reportedly forecast Ebola outbreaks in the DRC with 83 percent accuracy.
Summing up his position, Manyika said the industry's real contest isn't about who builds the most capable system first. "It's a race to get it right," he said.
RECENT STORIES
-
-
-
-
-