13 hours ago
- Business
- San Francisco Chronicle
S.F.'s largest employer says bots do ‘50%' of the work. Employees should be worried
As one of San Francisco's more outspoken tech leaders, Marc Benioff is no stranger to media firestorms, provoked by his unflinching stances on homelessness or surprise cheerleading for President Donald Trump.
The Salesforce CEO ignited a new controversy on Thursday, when he said during an episode of 'The Circuit with Emily Chang' that artificial intelligence does '30 to 50% of the work' at his cloud computing company.
Long a happy warrior for AI, Benioff tried to put a positive spin on its encroachment into his company's workforce culture. With bots taking over labor once performed by humans, he said, the human employees can 'move on to do higher value work.'
But given that Salesforce has cut 1,000 positions this year, Benioff's remark stung. Salesforce is extremely influential, both as the city's top employer and a proponent of returning to the office. Benioff is also a highly visible public figure. His championing of robot labor comes at a moment of eviscerating layoffs across the tech industry. Although most companies haven't blamed automation for these cuts, they coincide with a major industry push toward AI.
After Thursday's segment aired, immediately drawing more than 100,000 views, people pushed back on social media. A few self-identified Salesforce employees insisted the CEO had exaggerated, and that the human workers at Salesforce are too valuable to be replaced by software.
To industry observers, Benioff is just being realistic.
'There's no doubt that AI agents are replacing, and will replace, a substantial chunk of the workforce,' said Professor Saikat Chaudhuri, faculty director of the Management, Entrepreneurship & Technology Program at UC Berkeley's Haas School of Business.
Chaudhuri describes this moment as having a similar tectonic effect as the Internet Revolution, which caused pain for print media and brick-and-mortar stores, before opening a new world of possibilities. It became 'something that people had to acknowledge,' Chaudhuri said, making some jobs obsolete but creating new ones.
He offered an example from the travel sector. In the old days, people made appointments with human travel agents to plan their vacations, find flights and handle other logistics. By today's standards, the agents were expensive and inefficient, and became anachronisms once online booking took hold. At that point, consumers could plan trips on their own, by doing Internet searches and entering their information into web pages that had to be built by human engineers.
Looking ahead, the travel agent could be revived, but not in human form. The new iteration would be a concierge-style AI bot that listens to the consumer's requests and culls flight schedules instantaneously. With booking web pages no longer needed, the engineers are, theoretically, freed up for more creative pursuits.
Data gathered by economists at Stanford University corroborates that view, suggesting that Benioff, while bullish on AI deployment, isn't necessarily an outlier. Rather, he's pulling back the curtain on what's actually a widespread practice. Usage spiked over the last two quarters, with 40% of firms across the country now saying they use generative AI at work — up from 30% last December, according to a survey called 'The Labor Market Effects of Generative Artificial Intelligence.'
'It's increasing very rapidly, even surprisingly,' said Jon Hartley, a policy fellow at Stanford's Hoover Institution and lead author of the survey. Silicon Valley executives, in turn, have become more unapologetic about gambling in the technology. During a recent call with shareholders, Nvidia CEO Jensen Huang envisioned a future society with 'billions of robots.'
At some level, corporations have an incentive to put out these messages, said Jeffrey Hancock, a professor of communication at Stanford University and founding director of the Stanford Social Media Lab. Over the past few years many businesses have invested heavily in AI programs, and most are seeking some kind of return. They at least need to convey to shareholders that the bet was worthwhile.
Zeal from the C-suite doesn't necessarily mean the employees are on board. Some might feel intimidated and reject the technology altogether. Others will simply program the software to do their jobs, viewing it as a substitute rather than a helper. Many will find this new phase of work to be discouraging and dehumanizing.
Two schools of thought have emerged around artificial intelligence, Hancock said. The first is a 'pilot' mindset, which casts AI as a tool to augment performance and help people achieve their goals. Hancock describes the second mindset as that of a 'passenger,' or someone who feels a loss of power and control when interacting with AI.
By many measures, Benioff seems to fit in the first camp. He embraced a concept called 'agentic AI,' which refers to technology so sophisticated, it can accomplish tasks without human supervision. At its best, this class of AI should broaden possibilities for workers and create an opening for projects that previously would have been too burdensome. This is the AI that reads and summarizes your emails, manages your calendar, handles the customer service call with that person who wants to return shoes.
Benioff is so enamored of this model that he refers to Salesforce's AI bots as an army of digital 'agents.' According to Chang, he spent more than $20 million to license physicist Albert Einstein's likeness for Salesforce's AI branding. Salesforce seeks to have 1 billion active agents by the end of the year.
Whether Salesforce workers accept this new 'AgentForce' depends largely on how they interpret Benioff's attitude toward AI. And his remarks this week could serve as an inflection point.
Because the technology is so new and could so profoundly change labor, executives are uniquely positioned to shape the mentality at their companies, Hancock said. If Benioff portrays AI as productive and liberating, then his optimism will flow down from managers to lower-level employees. Hancock calls this a 'leadership cascade.'
On the flip side, if Salesforce employees construe Benioff's comments as a warning that he plans to lay off half the staff, then people will feel degraded and overwhelmed, as though they're being whip-sawed by new machinery.
'It's easier to imagine what will go away,' he said. It's much harder, at this juncture, to conceive the future benefits.
