Generative AI has become a top priority among businesses even though IT leaders are expressing concerns about potential ethical issues posed by the technology, according to a new Salesforce survey.

Sixty-seven percent of senior IT leaders surveyed said they will be prioritizing the technology over the next 18 months, and 33% claimed it would be their top priority, the survey said.

Despite its perceived benefits, however, respondents to the survey remain skeptical about some of the ethical challenges currently surrounding generative AI, in particular that its output could be biased or inaccurate.

Various forms of AI have been used by businesses for decades. Generative AI is the latest major development in the field. According to IDC, it is a form of artificial intelligence that uses unsupervised and semi-supervised algorithms to create new content from existing materials, such as text, audio, video, images and code.

The Salesforce survey, which asked 515 senior IT leaders in the US about their thoughts regarding generative AI, comes a week after CEO Marc Benioff told analysts that the growth of AI presents an opportunity for Salesforce. It also precedes the launch later this week of the company’s EinsteinGPT, which Benioff said will complement the company’s Einstein technology.

Launched in 2016, Salesforce Einstein is an integrated set of AI technologies that brings artificial intelligence into all Salesforce products, which the company says ultimately provides customers with more personalized and predictive experiences.

Although 33% of respondents to the Salesforce survey think the technology is already “over-hyped,” 57% said they believe that generative AI is a “game changer.” According to the report, better serving customers, helping to take advantage of data, and allowing organizations to operate more efficiently were cited as the top benefits of the technology by 87%, 80% and 79% of respondents respectively.

In addition, 79% of senior IT leaders said generative AI will help reduce team workload and thereby reduce burnout, and 77% believe that the technology will help their organization serve their customers faster. Seventy-five percent of respondents said generative AI helps their organization sell efficiently.

IT leaders express concerns about AI

However, as recent events have shown (Microsoft’s Bing chatbot recently professed its love to a New York Times reporter and told him to get a divorce), generative AI is not without its problems, a concern reflected by a significant percentage of the senior IT leaders surveyed by Salesforce. Seventy-nine percent of respondents believe that the technology has the potential to be a security risk, while 73% are concerned it could be biased, and 59% of respondents believe generative AI outputs are inaccurate.

Furthermore, 66% of respondents said their employees don’t have the skills to successfully leverage generative AI, while 60% believe the technology won’t integrate into their current tech stack, and 59% don’t have a unified data strategy to implement generative AI successfully.

Consequently, 99% of senior IT leaders surveyed believe their business must take measures to better equip themselves to successfully leverage the technology.

The importance of acknowledging the technical and ethical concerns regarding the implementation of AI was further highlighted by the fact that 83% of respondents to the survey think businesses must work together to ensure generative AI is used ethically.

In comments posted along with the announcement of the survey findings, Clara Shih, CEO of Salesforce’s Service Cloud, said that generative AI represents a change in how organizations across industries will analyze data, automate processes, and empower different departments to improve customer relationships. However, she also warned that the technology is not without new risks and challenges.

“Whether generating a tailored sales email or customer support chat response, an ethics-first approach grounded in trusted data and human-in-the-loop workflows is what will allow enterprises to safely and responsibly use generative AI to deliver against today’s growing customer expectations,” Shih said.

Artificial Intelligence, Enterprise Applications

The sixth annual report from Tech Talent Charter (TTC) has revealed that while companies in the UK are making progress toward improving diversity in their overall workforce, there is still a significant lack of diversity among senior technology leaders.

The not-for-profit charity, which focuses on tracking diversity in technology, compiled its report using data from 649 signatory companies, including Global, HP, Lloyds Banking Group, Nominet, PwC and CWJobs. The 210,245 employees included in the data set are estimated to represent around 16% of the UK’s technology workforce.

The Tech Talent Charter is free to sign — the only obligation signatories are required to meet is sharing their data with the charity when requested. It is only mandatory for signatories to share gender and ethnicity data but this year’s report represents the first time the TTC has started to track other aspects of diversity, including age, disability, sexual orientation, religion and neurodiversity.

The aim of collecting this data set is to try to understand what is actually happening at the coalface of diversity and inclusion in tech, since not being able to fill shortages in the tech talent market costs the UK economy about  £63 billion a year, according to Tech Talent Charter COO Lexie Papaspyrou.

Commenting on the report’s key takeaways,  Papaspyrou said that while it’s heartening to see that 28% of tech workers are gender minorities and 25% are from minority ethnic backgrounds, when those figures are compared to the percentage currently holding senior leadership positions, the drop-off is  alarming.

The data collected by TTC found that 22% of senior tech roles are held by gender minorities, a figure that is 6% lower when compared to tech roles overall, while ethnic diversity almost halves in senior roles, dropping from 25% to 13%.

There’s a pervasive idea that these figures just highlight the fact that a large percentage of women naturally leave the workforce at certain point to start families, but that doesn’t explain the drop-off experienced by people from an ethnicity minority background,  Papaspyrou said.

“There are no natural barriers that exist for ethnic diversity and senior roles in the same way that maybe you could argue exist with regards to gender,” she said. “There is a gendered societal problem that women are dropping out of the workforce because they need to take parental or career breaks, but that’s not the case when it comes to ethnicity.”

Papaspyrou added that she doesn’t believe it’s a coincidence that, for senior roles, the data for gender parity is more positive than the data for ethnic parity, given that the UK government has made gender pay gap reporting mandatory, but not ethnicity.

When it comes to D&I, data is key

One of the founding tenets of the Tech Talent Charter is the importance of data, so much so that if a company fails to provide TTC with the information required from them, they are removed as a signatory.

One of the ways the charity is helping organizations to have a better understanding of where they are on their D&I journey is through dynamic benchmarking, with a new tool that is freely available and allows companies to input their diversity figures and see how they compare to other organizations of the same size in their region and sector, Papaspyrou said.

“Those are the three areas we are constantly questioned about by companies,” she said. “They say, ‘I don’t know how to contextualize my figures because the publicly available ones are across all UK companies and I’m a small SME in the North East, so that isn’t relevant to me’.”

For TTC, while there is still an amount of churn regarding organizations that refuse to hand over their data, many companies are entering their fifth or sixth year of being signatories to the charter.

As a result, their support and willingness to provide more data has led to TTC being able to ask questions through eight diversity lenses, with neurodiversity emerging as a distinct area of interest, Papaspyrou said. Among current signatories, 53% are now measuring neurodiversity among employees, a figure that has doubled from last year.

Measurement of social mobility lags

However, Papaspyrou  was concerned about the data gap that exists regarding the measurement of social mobility, which lags far behind other areas including age, religion and orientation.

“We need to be looking at these intersections and social mobility is the one that falls across every single other lens and has such massive tangible impact on what works and what doesn’t work,” Papaspyrou said, adding that it was dismaying to see that lack of reporting, particularly in a year where the country is going through such economic hardship.

“When you look at the technology industry and where tech salaries are, in some cases three times the average UK salary, the fact that more organizations are not focusing on this is a big opportunity lost,” she said.

Looking forward to the next 12 months, Papaspyrou said TTC will be working with its signatories so that when the next data set is collected, organizations are ready to tell the truth, whether the data they have is good, bad, or unavailable.

There will be a lot of activity focused on how to improve progression at higher levels and removing the barriers that are stopping diverse employees from reaching those roles, Papaspyrou said.

“It seems like the entire business community here has really picked up on this idea that you can pack as many people into the tech workforce as soon as you want. And, while it’s great and you’re getting them in, if they’re not getting on, why are we doing it?” she said.

Diversity and Inclusion