The convergence of AI and Blockchain: what’s the deal? - Part IV

AI and Blockchain


By Francesco Corea, Ph.D.  |  May 24, 2018
Francesco Corea is a Complexity Scientist, Tech Investor, Data Strategist and AI Advisor


IV. Decentralized Intelligent Companies

There is plenty of landscapes of blockchain and cryptocurrencies startups out there. I am anyway only interested in those companies working at the intersection (or the convergence, as someone calls it) of AI and blockchain, which apparently are not that many. They are mainly concentrated in San Francisco area and London, but there are examples in New York, Australia, China, as well as some European countries.

They are indeed so few of them that is quite hard to classify them into clusters. I usually like to try to understand the underlying patterns and the type of impact/application certain groups of companies are having in the industry, but in this case is extremely difficult given the low number of data points so I will simply categorize them as follows:

  • Decentralized IntelligenceTraneAI (training AI in a decentralized way);Neureal (peer-to-peer AI supercomputing); SingularityNET (AI marketplace); Neuromation (synthetic datasets generation and algorithm training platform); AI Blockchain (multi-application intelligence); BurstIQ(healthcare data marketplace); AtMatrix (decentralized bots); OpenMinedproject (data marketplace to train machine learning locally); Synapse.ai(data and AI marketplace); Dopamine.ai (B2B AI monetization platform);Effect.ai (decentralized AI workforce and services marketplace);
     
  • Conversational PlatformGreen Running (home energy virtual assistant); Talla (chatbot); doc.ai (quantified biology and healthcare insights);
     
  • Prediction PlatformAugur (collective intelligence); Sharpe Capital(crowd-sourced sentiment predictions);
     
  • Intellectual PropertyLoci.io (IP discovery and mining);
     
  • Data provenance: KapeIQ (fraud detection on healthcare entities); Data Quarka (facts checking); Priops (data compliance); Signzy (KYC)
     
  • TradingEuklid (bitcoin investments); EthVentures (investments on digital tokens). For other (theoretical) applications in finance, see Lipton (2017);
     
  • InsuranceMutual.life (P2P insurance), Inari (general);
     
  • MiscellaneousSocial Coin (citizens’ reward systems); HealthyTail (pet analytics); Crowdz (e-commerce); DeepSee (media platform); ChainMind(cybersecurity).


A few general comments:

  • It looks interesting that many AI-blockchain companies have larger advisory board than teams. It might be an early sign that the convergence is not fully realized yet and there are more things we don’t understand that those ones we know;   
     
  • I am personally very excited to see the development of the first category (decentralized intelligence) but I also see a huge development in conversational and prediction platforms as well as intellectual property. I grouped other examples under “miscellaneous” because I don’t think at this stage they represent a specific category but rather only single attempt to match AI and blockchain;
     
  • Those companies are incredibly hard to evaluate. The websites are often cryptic enough to not really understand what they do and how (a bit paradoxical if you want to buy the blockchain transparency paradigm) and technology requires a high tech-education to be fully assessed. Cutting through the hype is a demanding task and this makes it very easy to be fooled. But let me give you a concrete example: ever heard of Magos AI? In the effort of researching companies for this post, I found myself reading several articles on this forecasting blockchain AI-driven platform company (WiredPrnewswire, etc.), which just did an ICO for over half a million dollars and that made great promises on its deliverables. The website didn’t work — weird, if you consider that they need to share material/information on the ICOs. But you know, it might happen. I made then an extra effort because I read it on Wired and I was curious to know more about it. I was able to find its co-founders, which I couldn’t find eventually on Linkedin. Weird again. Well, there are people who do not like socials, fair enough, especially if you consider that until three months ago there was no proof of the company existence whatsoever. Let me look into the rest of the team. Nothing even there, and no traceable indications of their previous experiences (except for the CTO master in analytics, that I found no proof of). I tried to then dig into the technology: white papers, blue papers, yellow papers, you name it. I only found reviews of them, no original copies. Final two steps: I don’t consider myself an expert in blockchain at all, but I read, a lot. And I also believe I am fairly knowledgeable when it comes to AI and what is happening in the industry. These guys claimed they created 5 different neural nets that could achieve the same accuracy in complex different domains than Libratus (or DeepStack) reached in Poker, but I never heard of them — very weird. Well, you know what? Maybe I could write them and meet them to understand. Their address points to the AXA office in Zurich. Ah.

After 5 minutes of research, I finally Google the two keywords: “Magos scam.” It seems these guys took the money and disappeared. They are surely building the 6 neural net somewhere, so stay tuned.

My point here is that exponential technologies are fantastic and can advance mankind, but as much as the benefits increase also the potential “negative convergence” increases exponentially. Stay alert.


Conclusion

Blockchain and AI are the two extreme sides of the technology spectrum: one fostering centralized intelligence on close data platforms, the other promoting decentralized applications in an open-data environment. However, if we find an intelligent way to make them working together, the total positive externalities could be amplified in a blink.

There are of course technical and ethical implications arising from the interaction between these two powerful technologies, as for example how do we edit (or even forget) data on a blockchain? Is an editable blockchain the solution? And is not an AI-blockchain pushing us to become data hoarder?

Honestly, I think the only thing we can do is keep experimenting.


Series: The convergence of AI and Blockchain: what’s the deal?

Part I: http://cognitiveworld.com/article/convergence-of-ai-and-blockchain-whats-the-deal-part-i

Part 2: http://cognitiveworld.com/article/convergence-of-ai-and-blockchain-whats-the-deal-part-ii

Part 3: http://cognitiveworld.com/article/convergence-of-ai-and-blockchain-whats-the-deal-part-iii

Part 4: http://cognitiveworld.com/article/convergence-of-ai-and-blockchain-whats-the-deal-part-iv

 

This series was originally published in December, 2017 on Medium and will appear in a forthcoming book edited by Spriger (2019).