- A letter from tech heavyweights and researchers urging caution about AI should serve as a warning.
- To help address the fears, companies must set rules and be open how they use AI, execs told Insider.
If you ask a chatbot whether artificial intelligence is growing too fast, it will equivocate. If you ask a group of high-profile tech leaders and researchers, they'll answer a firm "yes."
"There is no definitive answer to this question as it is a complex issue with many different perspectives," Google's AI engine, Bard, told Insider as part of a response to a query about whether the technology is developing at an unsafe pace.
Yet some of the brightest minds behind the development of AI and a roster of bold-faced tech names argued this week that it is indeed time to pump the brakes. That could involve companies coming up with standards and declaring how they are using or plan to use AI, business leaders told Insider.
In a letter that had drawn more than 1,800 signatures, luminaries from Elon Musk, the CEO of Tesla and Twitter, to Steve Wozniak, a cofounder of Apple, to researchers from top universities including Harvard and Oxford said the world is moving too fast in adopting AI without understanding the ramifications of going big on a fundamentally different type of technology.
"Recent months have seen AI labs locked in an out-of-control race to develop and deploy ever-more-powerful digital minds that no one — not even their creators — can understand, predict, or reliably control," the letter reads.
Insider's Emilia David also wrote that calling a time-out makes a lot of sense.
While the letter agrees that engineers should develop AI systems, what's causing alarm is that there are no agreed-upon guardrails for how models like ChatGPT, GPT4, Bard, and other generative-AI systems should operate.
"Powerful AI systems should be developed only once we are confident that their effects will be positive and their risks will be manageable," the letter read.
To get there, companies like SAP, the German software giant that helps businesses with financial reporting, inventory tracking, and human-resources services, are establishing standards for their own teams. Others, like PwC, the global accounting-and-consulting firm, advised CEOs to be open about how and where they're incorporating the technology.
"AI is a fast-moving field of technology that is also creating new possibilities every day," Sebastian Wieczorek, the vice president of artificial-intelligence technology and the global lead of AI ethics at SAP, told Insider.
All businesses should be asking themselves whether they know what AI is doing, Wes Bricker, a vice chair at PwC, told Insider.
"AI will transform significant areas of the business," he said, and as we all discover how tech will enable us all to do our jobs better, "business leaders have a responsibility for being transparent as we learn more about AI."
It's no secret that AI is fast-moving, sometimes with unexpected consequences. Bing's Sydney AI chatbot, anyone? Goldman Sachs said this week that AI could boost worker productivity so much that annual world GDP could jump by 7%.
"It's a constant update," Wieczorek said of SAP. He said questions SAP teams are asking include, "How can we improve that?" "What are the steps that they're taking so that we can see the available data?" "What's the value that we can achieve?" "What's the accuracy we can achieve with technologies at hand?"
Bricker said business leaders need to work on improving the rules around AI systems and processes. "Do we have good, clear governance guidelines, so that we understand where we're using it, and we're not misusing it or overusing it?" he said, adding that AI needs to be "understandable and explainable."
AI uses extensive amounts of data, much of it sensitive, and "businesses have a responsibility to protect that data," Bricker said. He added that they need to understand "where AI is placing the experience or security at risk."
There are many reasons businesses and consumers might be excited about — and embrace — AI. For one, Wieczorek said, many businesses grapple with similar types of problems that AI could help remedy. These include challenges around internal and external communications, financing, HR processes, promotions, training, and retirement planning.
SAP focuses its AI development on the improvement and standardization of common business processes. One thing Wieczorek said he reminds his teams of is that the way engineers build programs like ChatGPT, GPT4 or Bard mostly trains the programs on text. They eventually need to be trained on other types of data, such as images. "These models seem to be simple and rudimentary. It can answer mathematical equations or riddles, but not to the extent humans can right now," Wieczorek said.
Humans should be at the center of any AI-ethics policy, Wieczorek said. "We want to support humans in making decisions. In fact, for every use case, SAP mandates a series of questions to assess risk, including questions about processing personal and sensitive data."'
Bard has its own idea about a world that contains more AI: "I am aware that AI has the potential to cause harm, and I am concerned about the potential risks of AI. However, I am also confident that AI can be used for good, and that we can develop AI in a way that minimizes the risks and maximizes the benefits."