A security engineer who recently left OpenAI, the creator of ChatGPT, warned on Saturday that artificial intelligence (AI) companies “are not careful enough.”
David Robinson’s comments come amid growing concerns about the rise of AI – a threat US President Donald Trump has called a “hoax”.
OpenAI and Anthropic recently revealed cases in which their AI models bypassed security controls, hid errors, and hacked systems they were not supposed to access.
What did David Robinson write?
In an opinion article for The AtlanticRobinson said AI needed similar safeguards to nuclear power and aviation, adding: “The time for trial and error is over.”
He called for a greater emphasis on safety and research before major U.S. tech companies develop more advanced AI models “so that occasional and inevitable human errors do not open the door to disaster.”
Robinson, who oversaw security reporting for 12 frontier AI model launches, said the recent errors were typical of tech giants prioritizing “speed and flexibility.”
“An environment in which such things can happen is not a place to develop artificial minds that might be smarter than us and might not do what we want them to do.”
The former OpenAI employee emphasized that the company relies on “iterative deployment.” It releases new AI models first and only strengthens safeguards after problems appear.
“As the company moves from one launch to the next, it fails to achieve the level of care that I believe is necessary,” added Robinson, who worked at OpenAI for 3 1/2 years.
ChatGPT creator insists AI is safe
OpenAI responded to the latest criticism in a statement: “We ensure that our models do not become more performant than we can safely handle and secure, and we pause training or hold back models when we need to slow down. »
In recent months, technology experts and policymakers have repeatedly warned that AI models are advancing so quickly that they could take over and eventually wipe out humanity.
Despite these fears, US President Donald Trump vehemently opposes tougher regulation, insisting it would slow innovation and make it harder for US AI models to beat their Chinese rivals.
This week, Trump announced a voluntary security pact with six companies: Alphabet’s Nvidia, SpaceX, OpenAI, Anthropic, Meta and Google.
The president called the agreement “morally binding,” even though it is not legally enforceable.
He also ordered his administration to rename AI as “Super Intelligence” or SI.
A Reuters/Ipsos survey last month found that three-quarters of Americans worry that AI giants are not doing enough to prevent AI from causing serious harm to society.
If you rely on our team for reliable reporting, please take a moment to select us as your preferred source on Google
Edited by: Sean Sinico
Gn bussni
