Can We Tone Down the AI Doomerism?
President Donald Trump, in a speech to the U.N. General Assembly in late September, suggested changing the name of artificial intelligence to super intelligence, but there are plenty of people who don't think AI is all that super.
U.S. legislators already in office and candidates seeking office in November, from both political parties, have called for a moratorium on AI. People from the AI industry, including Anthropic CEO Dario Amodei, have also called for a slowing of AI research and development. And recently, a so-called whistleblower, former Anthropic engineer Jacob Coxon, said many top employees at some of the leading AI labs genuinely believe advanced AI systems could kill everyone by the end of the decade. Coxon also accused both Anthropic and OpenAI of acting irresponsibly in a rush to achieve artificial superintelligence.
And then there is a lot of backlash from ordinary citizens across the United States to the development of data centers, which are a necessary piece of the AI equation. Everyone wants Siri, Alexa, ChatGPT, and Claude to do as much as possible on their behalf, but no one wants the infrastructure needed to enable that to be built in their neighborhoods.
Yes, AI has become a political bugga-boo in the past few weeks, and it will likely play a huge role in the upcoming elections next month. Candidates for office will be judged not by where they stand on key issues like immigration, school choice, taxation, or the war in Iran but on whether they support or oppose data center construction, government control over AI, and tighter cybersecurity regulations. It's already appearing in many political attack ads.
In the CRM world, AI has come on fast and strong. In just a few years, AI—and its variations like generative AI and agentic AI—has permeated just about every form of marketing, sales, customer service, commerce, back-office, and management software available, and new applications of the technology are being rolled out almost daily. I can't remember the last time I received a press release that didn't mention the new AI capabilities that any particular vendor was rolling out in its latest software release. I often don't even mention in my product coverage that something is AI-enabled anymore because it is just assumed.
For CRM, it's far too late to put the AI genie back in the bottle. AI is ubiquitous, and many of the applications that use it now would cease to function properly if we were to pull back on AI.
That, to me, is the greater threat surrounding AI today. I’d be more worried about modern life and modern business getting on without AI than I would about a sort of "Terminator" doomsday scenario where the bots take up arms against us and wipe out all of humanity.
At the same time, I am willing to concede that AI, left completely unchecked, has the potential to do some real damage. The CRM industry has already taken steps to address this, as very few business AI systems are allowed to operate without some form of human-in-the-loop oversight, traceability, and data access restrictions.
The CRM industry and the larger tech world as a whole would do better to strengthen their work on AI guardrails than weakening the technology and America's position in the global AI arms race.
In his "Customer Experience" column this month, "5 Common Sense Elements for AI Safety in CX," (page 27), Ray Wang does a great job of laying out what's really at stake in the AI debate and what CRM vendors can do to better protect all our interests when it comes to AI. He suggests some very viable guardrails, including tighter access restrictions, periodic reviews by regulatory bodies, and even an ultimate kill switch. These are all worthy of consideration and can be explored right now—not five or 10 years from now.
President Trump and senior officials in his administration have shirked calls for the government to restrict AI. "I'm not going to stifle growth of something that will be bigger than the Industrial Revolution," Trump said.
Nowhere is that quite as applicable as it is in the CRM space, and forward-looking guardrails, not backward-thinking restrictions, are the way to make sure that AI doesn't go rogue and turn over all the details of my past Amazon purchases to fuel some supercomputer's plot to take over the world.
Leonard Klie is editor of CRM magazine. He can be reached at lklie@infotoday.com.