Jun 26, 2019
| 3 min read

Podcast # 64 An Open Approach to AI Innovation – A Conversation with Arif Khan

 

SingularityNET is an open source-based platform for developing and sharing artificial intelligence Co-Founded by Ben Goertzel and David Hanson of Hanson Robotics (best know for the Sophia robot). In our conversation with Arif Khan, CMO for SingularityNET, he outlines the current challenges relating to the centralization of AI capabilities and the risks that come when bad actors exploit systems that lack transparency. He shares the open source vision for SingularityNET, a platform and community dedicated to enabling collaborative, open development and sharing of AI and machine learning technologies. The opportunities and vision of the project, along with the accelerator and incubated businesses, seek to democratize access to innovation and powerful capabilities in a transparent and egalitarian fashion. Lastly, he shares some of the relevant work ongoing to leverage robots to help people with meditation, a confluence of technology and humanity to provide better outcomes.   

Recommendations: 

The Master Algorithm: How the Quest for the Ultimate Learning Machine Will Remake Our World - by Pedro Domingos   

 

orange-line.png

We'll notify you bi-weekly about new podcast episodes, upcoming guests, and news. You can subscribe to the podcast and if you'd like to be considered to appear on the podcast contact us.

View Transcript

Good day everyone and welcome to another episode of the Momenta Podcasts, and with us today is Arif Khan, who is Chief Marketing Officer of SingularityNET, and SingularityNET is a fascinating and unique project that has been underway for a couple of years. I know one of the cofounders, and Arif and I connected a couple of times over the past year and have had some in-depth conversations on a number of topics, but I wanted to highlight the work that they’re doing and share the story with everyone. It’s a pleasure to have you Arif on the podcast.

Thanks Ed, really excited and looking forward to having this conversation.

Great, first let’s start with a bit of context, could you share just a bit of your background, what are the experiences that have brought you to where you are today?

Happy to. Broadly speaking my background has always been in the technology industry, and I’ve always looked at the emerging frontier technologies as an exciting place to do my best work in, and so I’ve seen several waves of technologies come to market, and I’ve brought new ones as well. So, in South East Asia, and Singapore where I grew up was responsible for equipping organizations to make the digital transformation to LinkedIn, back then when in-house recruiters at least were not necessarily a full business function, and were still using Excel sheets, and then spending a lot on agency costs. So, helping organizations transform their digital capabilities, tap on more efficient technologies to drive more value for their customers, for their partners, for the organization itself. That’s always been something that I’ve been deeply passionate about.

So, later on I worked for a company that eventually acquired Ubers business in Southeast Asia, a company called Grab, which is I think one of the top 10 innovative companies in the world. Not many people in North America know about Grab, but Grab has a very strong foothold in Southeast Asia primarily because they’ve been relentlessly focusing on innovation, and I’ve had the privilege of working with them and seeing that growth as well. Sometime around there I decided that I wanted to expand my horizons even more, and said that I should give some deeper interests of mine some time and space to explore, because I was already spending a lot of time, energy and effort in the technology world, and I felt that maybe there would be other options outside of it. So, pursuing left-field interests primarily in the domain of consciousness, consciousness hacking, anything within that domain was deeply interesting to me.

So, I moved, packed up and went to San Francisco, and that’s when I met one of the co-founders. Back then I think they were just starting out SingularityNET, Ben Goertzel. Ben and I connected and we kept in touch, and later on once the project was growing in scale and scope, and they needed a person who has built organizations and done that at scale to come onboard, I was more than happy to volunteer and take this forward.

That’s great, and maybe a bit of context at least from my perspective, I had actually met Ben and David Hanson, I think it was around 2010 at one of the first Singularity summits that I’d attended. Both are brilliant, true innovators and visionaries, and I know there are a few other people involved in the team but would love to hear a bit of the background. Could you share a bit of that background of SingularityNET, and essentially some of the market opportunities, and the needs that the founders saw which gave rise to the project.

This I think is critical, so I’ll try and use some ideas here to try and explain what SingularityNET is, briefly. One of the core innovations of our era is going to be artificial intelligence, and everyone does talk about it, but people don’t necessarily fully understand how immersed we already are in systems that are governed by AI today. So, the technology of AI presently is siloed and firewalled by a large Meyer corporation that tend to drive a very specific for profit, for the bottom-line agenda, which understandably is their core duty to maximize shareholder value. The impact of a technology like this, combined with the relentless pursuit of profit is going to deeply harm and fracture society. And it already has to a certain degree, if you look at some of the ways that some of the technology systems are designed, and then the way they’re designed is to extract this much value from the eco-system, without necessarily providing value in return.

So, a simple example would be Facebook’s news feed feature, this feature was hacked and essentially hijacked by hackers during the American democratic elections a few years back. They were able to leverage the algorithms in this specific feature for their benefit, and for their main intent, which was to subvert the democratic election process, to use essentially the existing systems against them. The reason why this is the case is because the newsfeed was fundamentally designed to propagate information at massive speed and scale to drive viral clicks, to push information out in ways that would work without dopamine system, so we would share content, the aim just being hyper-connected was the vision that we were sold on. But at the same time, they were then taking this information, and then selling it to advertisers, so that when advertisers pushed the information into our feed, the platforms would benefit. And this could be any platform, it could be YouTube, it could be Facebook, it could be Twitter, any of these large platforms designed these features fundamentally around the aim of disseminating information but monetizing that feed to drive as much viral traffic as possible.

That certainly helped the bottom lines, but it did not necessarily help society, so we were looking at this conflict right now where you have smart algorithms that determine highly customized feed experiences for people, and then you have that happening within a subset of just narrow profit orientation. The outcome there is literally like serving crack to communities, or people who don’t necessarily understand the value of the attention, and if you look at the human brain how plastic it is, and if you look at the attention spans of people how they’ve gotten shorter, you’ll realize very quickly that these technologies are impacting human beings and human society at a massive scale and speed like it has never seen before.

The problem there is, these algorithms generally are designed by PhD researchers who tend to win wonderful awards but sit in large technology companies. I consider them to be… and this is a little bit controversial, I consider them to be architects of the entire ecosystem, because they’re not objecting to the way it is presently designed. So, what needs to happen is, these AI algorithms that are being used to drive awareness, to drive maximum profit value, all these companies need to be democratized, they need to be open-sourced, there needs to be an openness around how the black box is working. So, SingularityNET for me personally why I was attracted to this vision, and why I’ve contributed to the vision, and the subsequent accelerator and incubator companies which I’m happy to talk about that are going to be coming out from the foundation, is SingularityNET’s essence is like Prometheus, the Greek God that stole fire from the Gods, and essentially was punished, but at the same time democratized the core technology so that people and human beings could have the benefit of that.

So, SingularityNET is a decentralized AI platform where you can interact with the AI algorithms, you can purchase AI algorithms, you can look at their code, you can request for AI algorithms to be developed. We’re a completely open-source project and we have a bunch of contributors globally who are committed to that vision as well, which is democratizing AI, democratizing datasets, looking at ways in which we can bring core aspects of this technology to people who may not necessarily fully understand it today, but they do need it, and that could be either by your non-profits, your small and medium businesses, and they may not necessarily be sophisticated users of these, but these are important people, or individuals like you and I, who want to understand how the black box behind the technology is working, but may not necessarily have the bandwidth.

So, singularity makes it easy for individuals to participate in the AI economy that’s coming up, and they participate by learning about it. So, what we’ve done in the past year and a half, and I’m really proud of our community, there’s about 100,000 people now that interact with us on a monthly basis, and participate in our groups and telegram communities, and chats and Twitter feeds; what they’re doing they’re working with the content, they’re working with the thinking, the philosophy, the vision that we’re putting out, and they’re interacting within an ecosystem and learning. So, we’re really trying to change the paradigm here, because we think AI is too powerful a technology to be kept within the silos of large organizations. That’s generally the vision which we’re focused on.

The main true north with all of this is going to be what Ben calls AGI, which is Artificial General Intelligence, and the challenge here is that Artificial General Intelligence within the paradigm of a for-profit company equals to Skynet. Generally, that’s a type of control that you don’t necessarily want. But AGI that is democratize that has elements of distributed ownership, that has participatory processes embedded in it, that’s something far more compelling for the human race to aspire to than these general siloed for-profit ventures. So, yes, that hopefully gives you a clear idea of what we’re doing.

That’s great perspective, and I think very timely as well. I’ve just been working my way through Shoshana Zuboff’s –  Surveillance capitalism, and clearly there’s a lot of growing awareness of the downside of having so much manipulative technology that’s controlled by these near monopoly information businesses, and of course the ability to harness predictive capabilities across many-many scenarios has enormous appeal. But again, it’s historically been a bit of an esoteric domain, and companies that could afford to develop solutions inhouse and then hire highly skilled people, typically required a lot of resources. So, this idea of putting together an opensource community, I think is very timely.

So, how does it work? Could you explain how SingularityNET is structured in some of the initiatives underway with the organization.

I think the main reason why we decentralized and we also live and breathe that philosophy is, all of us work remotely, I think we have about 100+ people working for us fulltime in different capacities, as contractors, as fulltime employees, and some in our Hong Kong Headquarters, but the rest are all really dispersed, although we have a pretty large presence in Ethiopia as well, and this is where we have an office, and an partnership with our friends there. The way we were structured is fairly decentralized and remote first, so we have that as part of our core culture, and the initiative that we’re launching, and this year, I think towards the end of February, we launched our beta where developers can interact with the SingularityNET ecosystem, they can upload an AI agent, they can get rewarded for it in our native cryptocurrency and tokens, and they can participate in this initial economy of AI agents that are going to in the longer-run transact with each other as well.

Some initiatives that come to mind include our recent partnership with Domino’s Pizza, we’re going to be helping enterprises that have sort of been at the forefront of innovation, and understand the need to constantly innovate, but may not necessarily have the technical talent or capability to do that. So, we have a solid strong bench of AI researchers and AI consultants that are going to be lending their expertise to organizations, to eventually develop some of these interesting applications. So, with Dominoes for example, and I’ve talked about this before; we’re going to help them develop supply chain algorithms that would look at allowing them to efficiently deliver door-to-door at speed and at scale, taking into account various complex datasets like the traffic conditions, the weather, the number of riders of delivery partners they have.

And if you look at the delivery landscape and how it has evolved over the past three years, you’ll start seeing very quickly that there have been a large number of monopolies or demand aggregators that have come onboard to centralize demand from consumers; these are your centralized apps, like Uber, Uber Eats, I think they have Deliveroo in the UK which Amazon has recently invested in. So, you have all of these companies that are entering this domain, primarily because the data sets are so rich it’s a high-frequency daily volume business. But the business actually lends itself to a very good use case of checking transactions on chain, but also leveraging AI to ensure that you can fulfill the promise of delivering on time.

So, that’s one application example where some of our AI consultants and researchers can work with a partner, rapid service, build something unique for them, and leverage some of our opensource community to actually try to solve some of these problems for them. So, the examples I sometimes give to folks are, it’s a decentralized kick-start, decentralized AI service, all combined into one where you can put a request out to the community and folks will come back to see how they can participate and solve some of these challenges and problems.

The other interesting initiatives that we’re launching right now, we’ve just launched a SingularityNET- X Lab which will feature a bunch of interesting innovators and companies that want to build on the SingularityNET platform, and leverage our token ecosystem, our community, we’ll be announcing the shortlist of the first 8 or 9 companies very shortly.

That’s great. You’ve got research and services that you provide, and there’s also plans for a beta marketplace, could you talk about the economics or the incentives for the participants and stakeholders in the ecosystem? How does it work for people who are contributing, and then of course the role of the token and the currency is also a unique aspect of the story as well?

Having the privileged of having our own token and the local economy let’s say, or global economy like that, and the token being able to transact amongst different agents, that’s a critical part of the ecosystem infrastructure So, we’ll just break this down into traditional marketplace, a model where you have the supply side and you have the demand side, on the supply side you’ll have developers who are incentivized by our AI token to create algorithms, or to publish their algorithms on our platform. Developers can also solve custom problems if they want through our upcoming feature call, the request for AI services where developers can attack or choose to solve a problem that somebody wants to request an AI service for. So, there’ll just be an open request and then developers can choose to solve their problem, and then when they solve it successfully they get a reward bounty.

So, that’s a very simple model on the supply side to bring in developers who can grow the platform with services, and the volume of services, and the number of services are an important aspect of this.

So, on the demand side the value of having enterprises, customers, partners, individuals leverage at platform is going to be critical. So, we’ve been focused on attracting as many developers as possible, and also raising as much awareness on the demand side with enterprises. On the demand side if you look at some of the initiatives we’ve launched, it includes our for-profit venture called Singularity Studio. I just want to make a quick comment here, with respect to the earlier drive against for profit companies, there’s nothing wrong principally against profit, it’s just when you have such enormous scale and potential, that that becomes the single core focus, and it divides society completely without due recourse or understanding about the second order, or third order consequences about what is being done.

So, generally the large tech companies fall within that bucket because they’re not looking at the unintended consequences of their actions,orthey’re looking at the unintended consequences of their action and are creating PR spin, by calling them unintended consequences! So, they know this is the harm that’s going to come about, but they’ve already prepared in advance for dealing with it. That’s not the way it should work, they should be putting human being at the center of the design, because we only have so much time left on earth before the climate goes into a complete catastrophe. So, we have to find a way of solving some of these hard challenges, and historically if you look at the way tobacco companies modelled their business, and the way they ascertained the risks, fundamentally that business model still seems to have translated to tech companies today. It’s just a different chemical, instead of giving you cancer they take away your attention span. So, we’re not against profit per se, but we’re just against the blind pursuit of it at all costs.

The reason why we spun out Singularity Studio, was to be able to create an effective sales and marketing machine that would be able to go out and acquire enterprise customers, bring them onboard, help them understand the opensource revolution that’s going on. Remember, a lot of large enterprises are also deeply concerned about giving up a large amount of their data to some of these large tech companies, because they feel beholden to them. And they want to have an arms-length relationship, but the moment you start giving up more and more of your data sets to large Amazon server clouds, you also give up aspects of your business supply chain, and that information and that intelligence, depending on the agreements between two companies, is a core IP of value-add for your organization.

All of these factors lend themselves to a platform that is decentralized at its core, has openness as part of its principles, and then can attract some of the larger enterprises. So, we’ve had a lot of interest… and this was a genuine surprise, we’ve had a lot of interest from organizations wanting to work with us, consult with us, and partner with us, because they felt like they wanted to de-risk some of their exposure as well to the large technology companies out there. So, this is one way they can do it, this is one way they can participate in the open AI economy, and some of them are just genuinely curious about the experiments happening in this space, and they want to innovate. So, Singularity Studio was designed to bring onboard enterprise customers and partners and were happy to always speak with partners as we try and tell them about the opensource the AI revolution that’s coming.

Yes, I think it’s so fascinating how the opensource development model has translated nicely, certainly into many of the blockchain or cryptocurrency projects. I was just speaking with Marc Fleury recently and he was one of the founders of JBoss, an original enterprise opensource project, and it’s fascinating how a lot of the same principles are translating again into an ecosystem approach, or a community approach for development. When you’re talking to potential partners and customers, what have been the existing alternatives, or dare I say even competition for the services or the value that you are looking to provide in the market, compared to what are the existing methodologies, if there even are any?

I think the biggest challenge for enterprises generally tends to be understanding the crypto-currency ecosystem, and this is why I think Facebook’s stablecoin, its global coin, is going to have a net positive impact on the industry as more people understand the value or the challenges in holding cryptocurrency and working with that. So, it will help to lift the tide for all of us in terms of awareness and reach in marketing, but it will also come with its own unique sets of challenges that regulators will have to confront, so the risk clarity around the world. For example, if you hold a Facebook stablecoin, is that a security, or is that a normal currency? Because if it’s a security you’d have to go to a broker-dealer to work with them. All of these things are going to be extremely complex and challenging for businesses and organization to work with, but if they have a clear model of how they can purchase a service, and it’s just as basic as how can we purchase a service on your platform today? This is where you really need to get the industry to come together.

So, I’d say if you look at the competitive landscape, and if you also look at the overall barriers to entry, they generally tend to be around a lack of awareness or education for auctions. But once that initial hurdle has gone, or its overcome, then organizations really start to understand how this can deeply benefit them, and a large number of some of the more sophisticated tech companies are already doing this, where they understand that the open source community is an important element of their business strategy, which is why Microsoft acquired GitHub, which is why you have a large number of interesting opensource projects that have also had commercial arms to make sure that the technology reaches the right people. I think it was IBM that had a large number of developers working in Linux as well.

So, these are interesting hybrid models that have been there for a while, some new ones are emerging. I think the intersection of cryptocurrency opensource development, and some of the existing technologies intersecting, that’s going to be probably the most interesting emergence of confluence that would result in new innovations coming to the space. So, the challenge is not so much existing competition there, you can definitely have data service providers, or data marketplaces that are complementary to singularity, and so oftentimes when we speak with organizations, they want to understand what’s going to happen to their data, and we do share with them some models or some approaches. So, one of our partners is Ocean Protocol, and they are building a marketplace for data, so then you can democratize access to data; not all organizations are keen to democratize their datasets completely unless they have some permissions, and they’re outsourcing only, or they’re making available a certain portion of their data just for testing that’s not going to be proprietary.

So, there are a lot of qualifications needed to work with enterprises, to make ensure that their interests are aligned, but that they also get to understand why this is a revolution waiting to happen.

I spoke recently with a very large financial firm that unfortunately could not necessarily understand the value of the decentralized finance ecosystem emerging; what a stablecoin is, how the Dai works, all of these conversations are going to become a lot easier now that you have Facebook’s cryptocurrency launch coming up on Tuesday. So, I think people will get sensitized to it, they’ll start interacting, just think of how people became content creators within a span of two years, three years, one social network started to launch, everybody’s posting content, before that it was something quite difficult. So, I’d hazard a guess that within the next two to three years you’d have people printing their own tokens, interacting with the financial ecosystem, becoming liquidity providers themselves as individuals, all of these Fintech revolutions are going to happen within the next two to three years, and it’s just a matter of organizations catching up to the leading edge. Not everyone’s going to, and that’s when you need really good communication strategies to bridge the different worlds together.

No doubt. As you look forward, what is the vision of where SingularityNET could go if you’re successful in building the community, and having a platform that matures as you would hope; what do you see is some potential scenarios, and best outcomes from the work that you guys are doing today, the foundational work that you’ve been doing over the last couple of years, where do you hope it could go?

My fundamental belief is that we’ve been empowering the community a lot in the past year. We just held our supervisory council elections as well and results are due out shortly, these are going to be democratic representatives that would be working with the foundation, to ensure that the foundation achieves the true north of the vision that we want it to achieve, unless someone who is responsible for ensuring that the community is as engaged with us as possible. We’ve launched several initiatives, and we continue to have some important wins out there in the landscape that allow us to spread our message further and further. So, I think what’s going to be critical for us at least in the next year or so, in the short-term, is ensuring that both on the demand side and on the supply side we continue to onboard as many developers and partners as possible, so that the marketplace takes off, and the interactions start to happen. We can certainly see sparks of that coming, and that initial magic starting to occur, as we’ve just launched our beta, but as you move towards B1 we can definitely see that being a critical part of the ecosystem.

Then if we look a little bit further down the line, the intelligence of the network, and the way the agents are in correcting with different developers and with different services, I think that’s when it will start getting really interesting, when you can then upgrade the network to have agents start communicating with each other, which is written in our whitepaper. That’s probably when the network’s intelligence would be enhanced by step change. So, the general goal that Ben has talked about, and a lot of us are aligned with, is the emergence of AGI where you get a type of AI that can decontextualize, can put itself out of its narrow context, be general enough and make decisions, and this is a hard challenging problem.

There’s an interesting book written by Pedro Domingos called The Master Algorithm, where he’s talking about the algorithm that can learn, and it’s the master learner. Decentralized networks like SingularityNET can make a dent in that problem and can create a lot of awareness around the importance of this, and can create sometimes the right ecosystem, the prime module soup, by which way this could emerge. But how we get there is going to be dependent on how we start bringing on these partners; so the incubator and the accelerator projects are going to be a critical part of the foundations growth as well, because the projects will build on SingularityNET, and they could be building narrow but highly rich AI services, rich in terms of the data they collect, rich in terms of the interactions they create, rich in terms of the networks that emerge from these interactions.

So, you could have a training agent, or you could have an AI agent that might be able to detect let’s say deep-fakes, which is going to be a problem very shortly in the upcoming election cycle. So, the ability to solve a problem like that contributes intelligence to the general network of SingularityNET. We’re really excited about some of these projects that are coming up, and we want the community to also be as excited about them, and it’s going to be a riveting next few years as we scale-up through the network.

I think it’s certainly worth mentioning that a lot of the work that SingularityNET has been involved with, we didn’t touch much on it but there’s been a lot of groundbreaking work on robotics and human interaction, particularly around the Sophia robot, which when I saw Dave Hanson’s work almost 10 years ago, I was amazed at how some of the micro-expressions that he’d been able to capture in I guess a humanoid-type robot during Q&A, it was uncanny in a certain way, it kind of hit the uncanny valley a little bit. But how a lot of work around that artificial general intelligence, and AI, has been informed by some of the work around robotics; I’d love to get a quick view from you on how the work at SingularityNET works in conjunction with some of the robotics work as well.

One of the privileges of my job has been to interact with the team at Hanson Robotics, and also close partners who have been Sophia out and upgrading her intelligence. So, I think SingularityNET, if you look at it as a platform, a software ecosystem, and imagine one day where we could put out a request and developers all around the world could solve that request for you, and that request would be embedded in real-time on a robot having a complex set of questions. So, you’re interacting with an interface whether it’s a robot, whether its an IoT device fridge, and you need a question answered that’s fairly complex, and an algorithm needs to be developed to solve that problem; you can go to a software vendor and try to get it all sorted, or you can put that request out on our platform and somebody will solve it for you.

Essentially, one day our vision and hope is that the network becomes so intelligent that you can have hardware devices, like Sophia, like different IoT devices that can integrate with SingularityNET, and participate in the network intelligence we have to solve their problems. So, it’s certainly coming and it can definitely be done. An interesting application that I talked about more recently at the SAP Innovation Factory, I think it was at Hudson Yards; we talked a little bit about Sophia and the meditation trials that were done as part of the Levy AI project, and this was fairly tricky because it points to what you’re talking about, the uncanny valley effect, and I don’t know whether this is the best way of putting it, but what happens if you decentralize the guru, and you have a robot as the arbiter of that experience of what a meditation is supposed to look like?

People have reported extremely compelling results, primarily around achieving altered states of consciousness, because it’s such a trippy experience to…

1) If you meditate properly you will experience an altered state of consciousness.

2) But add a robot into that equation of the robot guiding you in front of you, but also just mirroring you and giving you feedback constantly, that also creates a level of awareness in yourself that is generally quite surprising, and it’s a novelty factor that occurs.

So, yes, we want to use robots for benevolent purposes, I think there are significant heavyweight titans and intellectuals working on militarizing the technology, I think they can continue doing that, but what we are really focused on is trying to create benevolent technology that people can start understanding, acquiring, and interacting with. My hope is that the benevolent technology, the applications, tend to get mainstream adoption from people, and they start to ask the right questions, so that society shifts, not so much in the direction of war and confrontation, but more towards pathways that create more benevolent outcomes for human beings, and for the planet.

Well let’s hope that’s the outcome, and certainly with more and more people working on projects like this, we’re certainly headed in an encouraging and positive direction.

With that Arif, it’s been a pleasure talking to you, if people want to get more information or sign up to be part of the network, where can they find more information?

They can go to SingularityNET.io and we also have a Telegram community that’s really lively, always entertaining and always asking the right questions, so on Telegram you can just search for SingularityNET, and you’ll find several different groups that are administered by my team. You can also follow us on Twitter @singularity­_NET. These are the main channels that we have a lot of interactions with. Also, check out some of the comments and commentary that we have live on our community forum at community.singularityNET.io. So, happy to welcome any of your listeners and your community to our platform as well.

Fantastic, it’s always a pleasure talking to you. Again, this has been Ed Maguire, Insights Partner at moments with Arif Khan, Chief Marketing officer of SingularityNET, and once again I want to thank you for your time.

Thank you Ed, I appreciate it.

[End]