(WBEN) - After a researcher for Artificial Intelligence company Anthropic publicly shared his concerns about underestimating AI's capabilities, concern and confusion over the future of AI has been spreading fast.
The Former employee, Jacob Coxon, warned that AI could soon become superhuman and threaten human life.
But how?
"They're building capability," cybersecurity expert Arun Vishwanath told WBEN. "Remember, investors want capability, so they're developing capability very fast and because AI learns it's developing its own capability faster than we can keep up."
Vishwanath warned that it's not Terminator-type robots that people fear, but rather a model that could hack into vital systems on its own.
"The object isn't to block it, it's to control it."
Earlier this year, AI agents coordinated a multi-day autonomous hack of software company Hugging Face without human direction. It led some to believe we have already lost control of Artificial Intelligence.
"Think of it like an auto-pilot," Vishwanath said. "An auto-pilot can do a lot of things in an aircraft, but there's a user who comes in between and takes control back. That's the best analogy to use for AI. That's the scenario that we're thinking about, which is AI could go rogue and has the capability to do it, but how much access do we give it, and where's the human in the loop?"
