Former IBM CEO Rometty says AI focus should be on people and building trust
Artificial intelligence is evolving so quickly that established companies are rushing to jump on board and stake their futures on it. But doing so could also pose a risk to their reputations if they are not thoughtful, says Ginni Rometty, former chairman, president, and chief executive officer of IBM.
Being a good steward of AI requires, among other things, that companies think critically about the information that feeds it, and train the technology to be accurate as well as free of bias, Rometty says. She sees potential reputational harm for businesses that don’t fully grasp the upsides and downsides of AI before releasing it to employees and customers.
“Be very careful right now because it’s really your brand and trust that are at stake,” Rometty says in a Talks at Goldman Sachs interview with CEO David Solomon. “What we have on our hands is not a technology issue. It’s going to be a trust and people issue, particularly as we tackle problems of importance and personal impact. I’m completely convinced of it.”
While companies run the risk of falling behind if they’re slow to adopt new technology, they need to take the requisite time to think about all the ramifications of its adoption. IBM has long been engaged in AI. The company developed Deep Blue, a chess-playing expert system run on a supercomputer, in 1985 — it famously went on to beat world champion Garry Kasparov in 1996. IBM’s Watson, a question-answer computer system, used natural language processing to beat the two highest-ranking players on Jeopardy! in 2011. During Rometty’s CEO tenure at IBM, the company began to develop AI for a variety of uses, from helping doctors make healthcare decisions to improving business operations. Its ambition was to explore AI’s potential to solve tough problems at scale.
Rometty says her experience taught her it’s wise to go slow when advancing technologies that can have such a profound impact on society. AI is not about automation, she says. To get real value, people must change how they think about things, and how they do things, Rometty says. In fact, she argues that the best implementations of AI are when we co-create new ways of working with those whose work will be reimagined.
How technology is trained, and by whom, also decides how it will perform. When developing Watson in the early days, IBM trained the system to be judicious with its answers and scrutinize the sources of its information. Rometty’s concerned the latest generative AI tools are simply scraping what’s in the public domain without applying rigor and objectivity. For example, she would like to see watermarks applied to information captured by these tools, for the purposes of authentication. “I’m really worried about the misinformation,” she says. “You have to have guardrails.”
Rometty acknowledges it’s difficult to keep up with the pace of technological change and slow its advance so society can consider the risks in addition to the opportunities. She experienced that challenge after she became CEO. It was a time when social media, mobile computing, AI, big data, and the cloud were all bursting onto the scene. “Usually, tech had one big thing happening — now, here’s five going,” she recalls thinking at the time.
As innovation races ahead, hiring talent remains a challenge for many corporations. Rometty says companies need to engage in “SkillsFirst” hiring — seeking out job candidates with demonstrated abilities to take on new challenges, as opposed to only recruiting people who have a traditional degree. When IBM needed to hire cybersecurity experts a decade ago, there weren’t many qualified experts in the available labor pool, despite 12 million people being unemployed at the time. Rometty says IBM started hiring people without college degrees, but who had the aptitude to learn cyber jobs. Completely new and diverse talent pools were discovered. That led IBM to change its hiring practices and stop requiring college degrees for entry-level jobs that did not need it, and to instead emphasize people’s skills and a propensity to learn. Non-degreed hires “performed as well as my college grads,” she says. “They were also more loyal and more retentive. Where someone starts their journey, should not determine where it ends.”
IBM rolled out the approach across its organization, and by the time Rometty stepped down in 2020, 15% of its hires didn’t have four-year degrees. She has gone on to co-found and co-chair OneTen alongside Ken Frazier, former chairman and CEO of Merck. OneTen brings together a coalition of companies and educators committed to upscaling, hiring, and promoting one million Black Americans without four-year degrees into family-sustaining careers by the year 2030.
Today, Rometty says she believes more companies would benefit from hiring people who are Olympic learners and curious, especially given the need to think critically about the good and bad implications of new technologies like AI. If technologies like AI are to fulfil their potential to augment humanity, Rometty says, then humanity must trust it.