Has Anthropic Just Derailed the AI Conversation?

Following Dario Amodei’s comments on the risks of AI the public discourse on the technology is in disarray. How should enterprises communicate in this environment?

For years, Silicon Valley has sold a utopian vision of AI. Tech leaders painted a picture of a world where “superintelligence” leads to “a new era of personal empowerment where people will have greater agency to improve the world in the directions they choose.” We were told we would live in a world where we wouldn’t need to work. Funded by a Universal Basic Income, we would be free to travel any path we chose. This month’s intervention by Dario Amodei, CEO of Anthropic, has provided something of a needle-scratch stop to that particular record. 

What did Dario Amodei say about AI?

Following recent reports of AI “swarms” escaping their labs and hacking websites, Amodei has called for the pace of AI development to slow down. The risks of AI are too great, Amodei tells us. We need to put the brakes on before it’s too late. It’s a sentiment that the likes of Musk and Altman have signalled their support for. 

It’s not my place to say whether Amodei is right or not. For every expert claiming that AI may destroy us all  in 10 years, there’s another saying that we’re all overreacting, or that AI leaders are just shooting for regulatory capture. 

What I can say, as a communications professional, is that Amodei isn’t doing anyone any favors. The manner in which he communicated has opened a can of worms, and led to more questions than answers. 

Amodei has hit the AI panic button, but we’re in a strange position where we can’t really be sure how far we can believe what he is saying, especially as all his erstwhile competitors seem to get in line so quickly. 

The state of public discourse on AI has never been worse.

How should enterprises communicate around AI?

Maybe this hand-grenade tossing approach to comms works for the AI giants, but for the vast majority of companies that are innovating with AI in day-to-day operations, this all makes their work much harder. 

How should they talk about what they’re doing when there is so much apocalyptic doom mongering going on around them?

My advice is to carry on regardless. There’s a major distinction between the frontier labs that are pushing the envelope on what’s possible with AI and enterprises using it to automate repetitive tasks, improve customer service, and drive incremental efficiencies. 

The media agenda is currently being dominated by big and tendentious claims about what AI might or might not do in the future. It’s essentially a lot of thought experiments and a world removed from how AI is actually being used to deliver value today. 

The public debate needs a healthy dose of realism from those enterprises currently bringing the technology into production. Steady, focused case studies of the ways AI is, for example, making logistics run smoother, or reducing friction in customer experience. And yes, also real world examples of how enterprises are applying robust governance and security policies to ensure that their AI applications are secure, safe, and compliant. 

Far from shying away from the AI debate, now’s the time for enterprises to steady the ship with grounded communications that are conjecture-free and backed up by real-world case studies.