In a few years, using blockchain will become as normal as using the internet. The same cannot be said about artificial intelligence, however, as it still faces a lot of problems related to the results it offers when analysing databases that are not of very good quality because it cannot sort the useless information from the database, according to Thomas Kulnigg, partner at Schoenherr Attorneys at Law, who talked to BR at MindChain.
“When you speak about the combination of AI and blockchain the point is that there are not so many applications where AI software uses a blockchain database, so there aren’t many regulations on that. However, from the experience with ICOs (initial coin offering) and blockchain-related transactions, regulators in Europe have clarified quite openly and transparently that they have a technologically neutral approach. They say we don’t care what technology you use if you meet the required regulations, like those in the capital markets.
Bitcoin or other cryptocurrencies are decentralised so there is no one that needs a permit. So it’s not considered money, but it is something of value and it can be taxed and regulated. That is the general approach and from a fiscal point of view – it is an income that can be taxed. In Austria the regulators state that it is ‘other income’ from the beginning so you have to pay taxes,” says Kulnigg.
Will blockchain be adopted on a wide scale?
“My long-term view is that blockchain will be as normal as the internet. Not a lot of people can explain how the internet technically works, but does anyone care? Do you care how a website or how Google works in the end? It is a technological solution and in this case, blockchain is a decentralized database. And the issue is that this it is something very new, but I remember that in the beginning of the ages of the internet and e-commerce nobody knew how to deal with that. And then regulators started to form certain views and the law started to adapt. I think that in a few years’ time there will just be blockchain-based applications and people won’t know what they can and cannot do, but I am positive that it will be very normal to use.
Maybe I’m wrong, but I am very skeptical that we will see the same technological shift in AI as we see with blockchain because AI lags so much behind in terms of the technological approach. Most of the many AI application do not work, are not that intelligent in the sense that we save time using them. What is even more problematic from our perspective is that you need so much time to make the AI learn.
For example, yesterday I met with an AI startup from Austria that developed an artificial intelligence software that extracts summaries from news articles that are 2 pages long, automatically. It has two or three bullet points so you can read 20 articles in one minute or at least their synopsis. But it took them two years to manually write 60,000 summaries of articles to make the AI learn how to make summaries! Two years! That is a problem, it is a lot of time, and if the data is not good, the result from AI also isn’t good.
The input data quality is very important. Take, for example, a bank. They have hundreds of thousands of customer files and I know from working with banks that the data quality is not always the best. If they use an AI to analyze data, the result will not be good. The same applies in the legal area, for instance we looked at AI applications that were saying how to make our contracts better by learning from our database, but in our database there were so many documents that are not correct from a quality perspective because you get comments from clients, you get comments from the other side and this has to be sorted out manually. Or you find another AI that does the sorting. This is the difficulty that I see. But the little steps are already taken so they will expand.”