Salesforce: Guardrails needed around generative AI at work

Leadership is worryingly overconfident about their workforce’s ability to leverage generative AI responsibly – this is according to a Salesforce survey of 4,000 workers in the US, the UK and Australia.

Credit: Salesforce’s Generative AI Snapshot Research Series.

It is clear that leadership are out of touch with employees. While workers are planning to use the technology (61%), and see real productivity benefits from generative AI – most notably that it could save them up to five hours a week (approximately one month a year) – employees are concerned about the security risks of generative AI (73%).

60% of employees told Salesforce that they don’t know how to keep data secure when using generative AI tools.

Other concerns employees have about generative AI are the biased (59%) and inaccurate (54%) outputs.

During a Gartner HR roundtable, senior director analyst Emily Rose McRae shared that these bias and inaccuracy concerns bring additional risk to organizations.

“If you’re using the publicly available tool, you may not be able to validate the accuracy of the information you receive. And that puts an organization at financial risk, reputational risk, legal risk in may cases”.

It’s time for guidelines and upskilling

It is clear that leaders need to not underestimate these risks – instead, they need to listen to employees and figure out the best path to follow.

While some businesses – most notably big banks and tech giants – have restricted or banned outright the use of generative AI at work, given the productivity benefits that might not be the most sensible approach long-term.

Instead, it is time for organizations, and specifically HR teams, to figure out the “best path to follow” for them, according to McRae.

Salesforce’s data shows that a good place for employers to start is with guidelines for AI use (58% of employees agree), as well as enhanced security measures (59%).

Talking about the findings of the report, Salesforce’s chief ethical and humane use officer Paula Goldman shared: “Generative AI has the potential to help businesses connect with their audiences in new, more personalized ways.

“As companies embrace this technology, businesses need to ensure that there are ethical guidelines and guardrails in place for safe and secure development and use of generative AI.”

Another key part of HR and organizations successfully using generative AI in an ethical manner is technology upskilling.

The UK segment of the Salesforce survey found that 62% of employees lack the digital skills to use generative AI safely and accurately. 67% said they were concerned about the AI skills of their colleagues.

While 55% said they wanted opportunities for generative AI training, 79% said employers provide none of this. This is a concern as learning providers like Udemy and Coursera have specific courses aimed at generative AI and general automation.

Salesforce CEO for UK&I Zahra Bahrolouloumi commented: “AI has the power to transform how work gets done, but it is only as good as the data it is trained on. Without high-quality, trusted data, it becomes ‘garbage in, garbage out.’

“To stay competitive and drive long-term transformation for their customers and business, organizations need highly skilled people to use trusted data sources and ensure sensitive data is kept secure.

“The need to prioritize upskilling at every level has never been more clear nor urgent.”