Arm CEO Rene Haas has raised concerns that humans could eventually lose control over artificial intelligence (AI) systems if there are not sufficient safeguards or override capabilities put in place.
Speaking to Bloomberg, Haas admitted he worries about humans "losing capability’ over the machines at some point in the future.
"The thing I worry about most is humans losing capability," he told the publication.
“You need some override, some backdoor, some way that the system can be shut down."
The rapid acceleration of the generative AI space has prompted concerns among portions of the global technology sector in recent months.
In March, a host of industry executives, including Apple co-founder Steve Wosniak and Tesla CEO Elon Musk signed an open letter calling for a six-month pause on training AI systems more powerful than OpenAI’s GPT-4.
The letter claimed “human-competitive intelligence can pose profound risks to society and humanity”, and outlined a list of the “minimum “ level of safeguarding precautions that should be implemented.
These precautionary measures include setting up new regulatory authorities dedicated to AI, oversight over the most advanced AI systems, liabilities for harm caused by AI, and more public funding for technical AI safety research.
A second open letter issued by the Center for AI Safety with a far more streamlined message was signed by notable names such as OpenAI CEO Sam Altman, Google DeepMind CEO Demis Hassabis, Anthropic CEO Dario Amodei, and Microsoft co-founder Bill Gates.
Make the most of the opportunities GenAI offers
“Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war,” the letter read.
Both letters drew some criticism, with critics arguing the letters distract from the actual near-term harms AI systems already pose to society by focusing on the more abstract, long-term consequences they may incur.
The rush to apply generative AI models across society has prompted concerns around the potential to exacerbate pre-existing biases, amplification of social inequalities, and whether the use of certain generative AI tools could infringe intellectual property (IP) and copyright laws.
The European Union (EU) recently announced it has reached a provisional agreement on its own regulatory framework for AI systems, with measures in place to guarantee the transparency safety of ‘high impact’ foundation models and restrictions around using AI in biometric surveillance systems.
The agreement came after a protracted period of negotiations, slowed down by concerns from industry leaders that innovation may be inhibited by cumbersome regulatory red tape.
Get the ITPro. daily newsletter
Receive our latest news, industry updates, featured resources and more. Sign up today to receive our FREE report on AI cyber crime & security - newly updated for 2023.
Solomon Klappholz is a Staff Writer at ITPro. He has experience writing about the technologies that facilitate industrial manufacturing which led to him developing a particular interest in IT regulation, industrial infrastructure applications, and machine learning.