In his 5,000-word Techno-Optimist Manifesto published on Monday, billionaire venture capitalist Marc Andreessen celebrates technology’s unbounded potential to drive economic, scientific, and social growth. “There is no material problem that cannot be solved by technology,” he writes. Slowing down artificial intelligence, he argues, would be tantamount to murder: “We believe any deceleration of AI will cost lives. Deaths that were preventable by the AI that was prevented from existing is a form of murder.”  

I share his optimism about technology, but not blindly. I am in what Andreessen and much of Big Tech call the “decel” – or deceleration – camp. I’m not the only one convinced that our lives depend on it, though I’m not accusing anyone of murder. 

Andreessen, one of the world’s biggest tech investors, is seeking huge financial returns. He has every incentive to create a sense that it’s too late to do anything about AI, so just trust entrepreneurs and the marketplace to work it out. 

But why should we give a hall pass to those who will benefit most financially from AI, without requiring they also keep us truly safe. Would we be okay with new types of air travel or drugs being rushed to market because they are awesome even if they aren’t safe? No. In that sense, Andreessen & Co.’s insistence that no holds be barred as they advance AI  more like reckless radicalism than optimism. 

I’m for progress. I’m pro-business. I’m pro-innovation. And I believe in companies' ability to come up with solutions if they are actually held accountable for creating them. If you see to it that companies can’t do business until they meet credible safety criteria, you know what? They’ll come up with answers real fast. 

I ushered corporate social responsibility into Nike, working with internal and external stakeholders to improve labor conditions. There the path to progress was formed with an aperture wide enough to take in impacts and consequences of all kinds – on the products, the profits, and the people. The Industry found a common point on the compass as our shared goal, and addressed the issues. And companies like Nike grew as they pursued growth and responsibility at the same time. 

From a regulatory viewpoint, AI is arguably the toughest challenge humanity may ever face, but it isn’t impossible to solve. 

Andreessen’s breezy optimism notwithstanding, we’ve seen the immense havoc a go-fast-and-break-things attitude toward technology can wreak. Early in its global ascent, social media showed signs of the kind of damage that its race for attention could inflict but those pointing this out were ignored. Tech leaders didn’t explicitly set out to harm teenage girls or undermine elections – that was collateral damage in the pursuit of growth.

I am not hearing any tech leaders say machine capabilities won’t surpass that of humans. They are debating when it will happen. So if we know it’s going to happen, let’s decide what the solutions are, what needs to get done and when the possibility to deploy them disappears. Right now there’s no plan. That’s the real problem. 

You don’t have to grasp AI and all its complexities to understand that we don’t know how to stop the car from driving us all off the cliff. Responsible tech leaders have raised the caution flag repeatedly in the media and during their high-profile briefing at the White House on risks of AI.  They’ve signed on to voluntary measures to safeguard the technology, but three months later, what concrete, measurable actions have been taken? It seems to me the car is still speeding toward the cliff. 

For one simple reason. Companies don’t have to do anything. Voluntary is just that - voluntary.  Until governments impose a deadline and create a system for reporting actions and progress, we’re all in the dark. That would be fine, except we’re the ones in the car.

Thankfully, it’s not too late. We have a rare -- if fleeting -- opportunity to pump the brakes and act before these AI-driven tools are so entrenched in daily operations that their dangers are normalized and what’s unleashed can’t be contained, just as we’ve seen with social media. We won’t have the chance to retrofit this industry. But we can incentivize and reward ‌AI winners for doing it right, not being fastest or biggest. This is not easy. Open Source AI, whose ChatGPT sparked the AI mania a little over a year ago, is charging ahead with no constraints. Competition among the world’s tech giants, from California to China, is mounting, and governments like the UAE want to get in the game.

While this may sound like great creativity being unleashed, the creators have no idea what exactly it is they are unleashing. Are you ok with that? Governments aren’t going to figure this out but they can require companies to come up with answers – and do so on a tight deadline. 

Solutions are never black and white, good and bad. But that unfortunately is how the AI dilemma is being framed by Marc Andreessen and others in the so-called e/acc community – short for “effective accelerationism.” Let’s all be honest and recognize that artificial intelligence is both good and bad, dangerous and full of amazing possibilities, and figure out how to make this a win / win for all humans, not just a small group of tech bros.  

There will be no redos. Techno-radicalism needs a reality check in the form of clear deadlines for companies to deliver on the voluntary commitments they have made. And if they miss that deadline, their license to operate should be revoked. We need mandated collaboration to create industry norms together now, not later, so we can all benefit from the bright future of AI.