Salesforce CEO Marc Benioff has entered the escalating discussion around artificial intelligence (AI) safety, warning tech giants to steer the technology responsibly even as automated systems grow increasingly capable. Speaking to CNBC on the sidelines of Salesforce’s flagship Dreamforce conference in San Francisco, Benioff pointed to the risks and dangers that AI have but said that the onus rests on developers to ensure its responsible creation and use. He compared AI with social media platforms.“A lot of companies got hurt, a lot of individuals got hurt through social media. We don’t want that to happen in AI,” Benioff said. The Salesforce founder’s comments arrive amid intense friction over frontier model safety, triggered by Anthropic head Dario Amodei’s recent essay urging developers to cool their deployment speed so security guardrails have time to mature.Benioff has long pressured Silicon Valley to prioritise civic welfare. An outspoken critic of social networks, he famously labeled Facebook “the new cigarettes” in 2018 while demanding aggressive federal oversight.
‘Responsible AI development needed’
Benioff did not criticise AI companies or endorsed calls to slow AI development, arguing that they have a responsibility to consider the consequences.“We want AI to take care of these things and be held responsible, and take care of these actions and be ethical. The heart of ethics is responsibility. We really look at a light on our brethren here in San Francisco and all over the world and say, ‘We want you to be ethical in your actions and how you take care of your communities and the world,’ because this technology is very powerful,” Benioff said.His comments came days after Microsoft published a draft of codes that it follows when developing AI.“We believe it is more important than ever to create safe and reliable AI in service of people, and to be transparent about how we go about it,” the company said.“We think this is a common sense and practical approach to making AI safe, secure, and in service of humanity. The Code is designed to ensure MAI models will never resist human interruption, correction, or shutdown. That they will not widen their own scope, take on goals no human has given them, or hide their reasoning from the people auditing them. There are Absolute Constraints, things the models should never do, covering areas like weapons of mass harm, child safety, and harmful manipulation at scale. But at the same time, it sets defaults that mean it should be both helpful and safe,” it added.
Salesforce has deep commercial ties with frontier labs
Benioff’s caution is particularly significant because Salesforce remains deeply connected with prominent machine learning developers. Last month, Salesforce rolled out Claudeforce, an enterprise integration allowing corporate clients to connect Anthropic’s Claude assistant directly to corporate customer records to draft communications and refresh account entries.