Time to Prepare for AI Superhuman
AI is controversial. However, its advantages are still tangible despite technology's risks, the experts claim.
In a new Open AI blog post called “Governance of superintelligence”, Sam Altman, Greg Brockman, and Ilya Sutskever argue that it is time to consider ways of regulating the industry in order to avoid severe problems in the future.
According to them, AI systems may take the reins in most domains in the next decade, be as prolific as modern-day corporations, and beat similar technology from the past. However, even though the advantages include what the authors describe as a “dramatically more prosperous future,” the disadvantages are that there is a highly likely risk involved. This is why it is crucial to tackle the threats of AI technology and treat superintelligence carefully.
The authors propose that effective solutions necessitate a certain level of coordination among the key development endeavors. For example, significant governments worldwide could set up a project that many current efforts could become part of. Or, for example, people could collectively agree that AI grows only at a specific yearly rate while also ensuring that individual companies will be held to an extremely high standard of acting responsibly.
There must also be the technical capability to make a superintelligence safe, an issue the researchers are working on.
They add that while not everything requires strong oversight, control is necessary when it comes to more powerful systems.
“We believe people around the world should democratically decide on the bounds and defaults for AI systems,” the writers add.
The authors also argue that technology is highly beneficial and could dramatically improve people’s lives worldwide and solve different problems. Moreover, they think that the economic growth and increase in quality of life will be nothing short of astonishing.
Accordingly, there’s no way of banning technology development due to its tremendous upsides and the cost of building it decreasing each year, which is why “we have to get it right.”
Previously, GNcrypto covered advancements in regulating artificial intelligence in the EU.
