Marc Benioff at the AI for Good Summit
This is not the greatest technology of our lifetime; this is the greatest technology of any lifetime – and it is going to have a dramatic impact. So we have to decide – are we going to have AI for good?
A handy summation of the underlying business of the AI for Good Summit which has been taking place this week in Geneva. The quote comes from Salesforce CEO Marc Benioff, a tech sector CEO who has spoken out on ethical stances on a wide number of topics across the decades, from sexual equality thru environmental concerns to LGBTQ+ representation.
But AI marks a whole new level in the tech and ethics debate, he suggests, one with massive societal and geo-political implications. This is a moment of responsibility, he declares:
The key is this – are [AI] companies going to be held responsible for the technology that they’re building? We all saw social media companies get a pass. Well, is this going to be social media 2.0, where companies can build whatever they want, deliver whatever technology they want?
Noting that he’s speaking at an event called AI for Good, he goes on to argue:
I think that is the point we’re at right now. We’ve seen, looking back at social media, the huge number of children that have been hurt, a huge amount of damage done to society, but where have been the consequences for those companies? That is the warm-up act for AI. So, is this social media 2.0, or are we ready to step up in our industry to another level of responsibility?
Action now
This isn’t the first time Benioff has aired these concerns. Back in January he used a panel session at the World Economic Forum annual meeting in Davos to rail against AI tech that was found to be encouraging teens to self-harm and kill themselves. At that time, he warned:
These tech companies will not be held responsible for the damage that they are basically doing to our families, just as the social media companies have not been held responsible for the damage that they did…These US tech companies, they hate regulation.
So what role does Benioff see for himself in this pivotal moment, heading up as he does a global enterprise tech firm that has bet the farm on AI? One reason for his presence at the Geneva summit this week is his new role as co-Chair of the International Telecommunications Union’s new AI for Good Global Commission. This brings together more than 40 founding members, including heads of state and government, industry CEOs – tech and non-tech- and heads of UN agencies. Notable tech sector names involved, apart from Benioff himself, include Amazon’s Andy Jassy, NVIDIA’s Jensen Huang, Anthropic’s Jack Clark, Microsoft’s Brad Smith, Cohere’s Aiden Gomez, and Accenture’s Julie Sweet.
Building on the foundation of the multi-stakeholder ITU/UNESCO Broadband Commission for Sustainable Development, which helped shape global priorities for connectivity, digital inclusion and economic development, the new commission has a mandate built on three pillars:
-
AI Trust, in the form of responsible and trustworthy AI systems.
-
AI Access in the form of infrastructure, skills, and readiness for wider adoption.
-
AI Impact, accelerating AI applications for health, education, food security, and disaster response.
Benioff’s fellow co-Chair is Rwandan President Paul Kagame who pitches:
One thing is certain: technology is supposed to be a force for good, and we have a responsibility to use it accordingly. Let us work together to reduce inequality, and allow more and more of our citizens to benefit from the good AI can deliver to all of us.
For his part Benioff says:
I think that President Kagame and I agree that it’s critical that we must have a multi-stakeholder dialog. Only through a multi-stakeholder dialog are we going to get to the point that we can have some understanding of how to move things forward.
He adds:
We have put together a really unusual group of people, not only the Excellencies, but also leaders from non-governmental organizations, governmental organizations, religious organizations, business, all aspects of civil society, represented in this group. I’m sure that no group can be put together that would be totally comprehensive, but we’ve done our best to make sure that it is as inclusive as possible, and that we can have many voices.
Capitalism kicks in
Coming at the issue from a tech sector point of view, Benioff states:
We’re building technology that can build itself, we’re building technology that is going to start to govern itself, and in some ways it’s possible that humans could lose control of this technology if we don’t set up the liability now, the fundamental liability structure, the responsibility structure. Where is the ethical framework? How can you have ethics without responsibility?…If we do not have responsibility, how will we have the framework of ethics and morality in regard to these agents that are self-creating? This technology is self-creating, governing itself, managing itself, developing itself. It will have its ability to operate without us, so how will it have values? Values will bring value, so, what values are we actually creating?
But in the tech ‘Wild West’, does moral value take second place to commercial value as it so often has in the past? Benioff is an unashamed capitalist CEO of a publicly-quoted company with shareholder responsibilities, and he admits that “a lot of people are about to make a lot of money” out of AI:
We just saw the SpaceX IPO, we’re going to see the OpenAI IPO, we are going to see the Anthropic IPO, and there is a ton of venture capital that is flowing into this industry, so there is a huge set of returns. Don’t forget these are products, each and every one of them is a product built by a company. It’s not some kind of technology that’s being generated out of some kind of free framework. The primary consumption of AI today is through packaged software…The models themselves rarely run independently as their own value proposition…But as these companies grow, my company, every company, needs to be held responsible for what we are building. We are building products that will impact society in a way and a level that has never impacted society before.
Commercial interest can be balanced with values and companies can be held to account. It happens in other sectors, Benioff points out:
There’s lots of pharmaceutical companies here in Switzerland. If they build drugs that harm humans, are there already any frameworks in place that held those companies responsible? If these kinds of frameworks are already in place, why are they not being put in place at that kind of level of vault for the rest of the technology industry?
But the reality is that they are not – yet. The rise of the major frontier models has been powered by wholesale theft, an industry built on stolen data, as Benioff acknowledges:
It’s amazing, isn’t it? Data was stolen, and the companies were not held responsible.
Actually a few have been held responsible, such as Anthropic, fined $1.5 billion last year following a lawsuit by a group of authors. Other cases are rumbling through the court system.
Interestingly, Salesforce is a heavy backer of Anthropic, to the tune of its investment equating to around one percent of the firm’s putative valuation.
Anthropic has also been the target of direct regulatory action by the US Government of late, with its Mythos and Fable models slapped with an export ban that resulted in a worldwide withdrawal of the tech, since reversed. Benioff approves of what happened, it seems:
I think that the US Government needs to be more aggressive in taking action. They saw it as a weapon, they shut it down. You know, it’s very interesting. Another tech CEO [Amazon’s Andy Jassy] made a call to the US Government Secretary, the Secretary got incredibly anxious, called another Secretary, and within a couple hours, it had an export control, which means [Anthropic] had to shut it down because they viewed it as a weapon. And that was the first time that we had seen an aggressive action.
I felt like I was living out AI 2027 [a research initiative from The AI Futures Project that looks at developing hypothetical scenarios for AI development] – here’s the government getting involved for the first time, taking a somewhat aggressive action against the model company. But then I said, ‘Good, that is great’. In Europe that created a lot of anxiety because all of a sudden, European companies, especially CEOs, said, ‘Wait, does that mean the US Government can stop this technology or stop the technology that I’m running my company on?’ , rather than looking at it as the US government just stepped in to stop a potential weapon that could harm the world. So I felt good about it. I felt good because I thought governments need to hold the company responsible for what they are building.
He concludes:
Technology is never good or bad. It’s what we do with the technology that matters, and that has never been more true than in regards to this technology.
My take
Are we going to just let this be social media 2.0, or are we willing to step in and create a new framework for technology governance?
The presentations at this week’s AI for Good Summit have been impressive with all the the right sort of declarations made on all the usual talking points around ethics and the need for action. But the question is what happens when the private jets head off from Geneva and touch down back home?
All too often throughout tech sector history, standard bodies and organizations have gotten together every six months or so, issued some positive affirmations of collaboration for the common good, posed for the obligatory photo shoot, then gone home and started to fight like a bag of cats all over again as vested commercial and national interests kick in.
Will this new commission be any different? Benioff makes the point:
This is not a regulatory body. This is not a rule-making body. This is a conversation, and I hope that it will be a conversation that helps to create a higher level of enlightenment of not only where we’re going, but how we can go someplace that can be more for good.
Well, conversations are always good – jaw, jaw, not war, war etc.
And the inclusion of civil society representation and non-tech business leaders alongside the tech moguls and the political bigwigs looks to be an interesting differentiator.
Will that make a difference? Only time will tell, of course, but the stakes are high as Benioff reminds us:
We’re going to see things in our lifetime that we’ve only seen in movies and science fiction. We’ll be seeing it come to life in an incredible way, and as we want that to happen, we want it to come through with AI for good.
