Legislating on safe artificial intelligence could give Malta the opportunity to show other countries that developing AI and AGI should not be done without keeping human interest in line, says Hyperion CEO and co-founder Stav Zilbershtein.
Is artificial intelligence (AI) a utopia or dystopia?
That’s the question everyone wants to know the answer for. I think in the short-term we won’t even notice how AI will take over more and more of our daily lives. It already does now. Each time you search for something in Google you are putting deep mind (Google’s AI) to work. We are constantly interacting with AI on our mobile phones when we use Siri or Alexa. The more we interact with it, the more it learns and improves.
In the short-term these so called narrow purpose AI – like the systems that play games, offer better songs on Spotify or recommend movies on Netflix – will expand into the automotive sector, such as self-driving cars that communicate as one network in real time, and start analysing medical test results and slowly even performing operations.
The long-term implications are unknown. First there is the actual economic challenge that will remove enormous sectors of common labour like drivers, cahiers and even cleaners and caretakers from the market, replacing them with automatic and almost errorless performers that never sleep and can work 24/7.
But even if we find a way to balance the economy and enable people to keep living in a good standard, there is the race for Artificial General Intelligence (AGI). This is the aspiration of at least 50 public groups – ones that are known – and dozens of others that are in stealth mode to create a human-like artificial intelligence. This doesn’t only mean that the system can solve problems like humans do, but also that this AI is aware of itself and constantly improves itself. It writes its own code and constantly tests its improvement.
The problem with such a system is that because it can use an ever-growing computing power it will surely surpass human intelligence. What would such a system do next? Is it safe enough to develop AGI without thinking about security? All these questions seems not to matter as this kind of system will come into being faster than we imagine.
No one really knows what the consequences of such a system will be. It could be the greatest advancement of mankind or just as well it could be its final invention that will put an end to mankind as we know it.
Scientists that work on such projects vary in their opinions, from sheer optimism to cautiousness. But since there is no regulation on developing safe AGI systems, our fate is in the hands of these companies that work on such projects and we can all hope the sheer optimism will turn into our reality.
Automation carries with it a negative baggage, especially with the perceived threat of loss of employment. But what are its benefits, especially within the iGaming and fintech industry?
Coming back to 2019 after a broader perspective in the short-term, we already see that AI is set to improve on identifying problematic behavioural patterns within players. The system can quickly flag when players change their behaviour for the worse and prompts the operators to take action in accordance to the responsible gaming regulations.
Other than that, I do believe that we will see a more user-based experience. An AI can segment patterns and users better than us humans. Our A-B tests are based on a broad observation and asking simple questions.
AI can constantly look at patterns and for example fragment what we will consider as one market (UK players) for example into many sub groups of which different content and games and calls to action can be used to engage with them in a far more efficient way.
For that to happen, companies need to first identify where the pools of big data are gathering and then think about what questions they want answered and which challenges are too complex for a human to comprehend because of the ever-increasing data volumes and start working with an AI to tackle these questions.
Does automation free up employees to do more creative thinking and strategy?
Yes and no. On the one hand people will become partners with AI. But this slot will be reserved to specific experts and much less of them will be needed to operate parts of the system. So divisions of 10 people could easily be narrowed to one expert assisted with an AI.
AI provides constant insights and suggestions and the human partner that works side by side with it can implement and learn what to improve
AI provides constant insights and suggestions and the human partner that works side by side with it can implement and learn what to improve. This means that for some people more creative work will definitely take place but many others will need to find their placement in the market. It might be that this kind of pressure will drive people to acquire more exotic technology skills by knowing that simple jobs are fading out. Is there a sustainable solution for a large scale of such occurrence? In my opinion not yet with our current economic system. We will need to rethink it. Perhaps AI will assist us with that task one day as it seems now that we can’t come up with something reasonable.
What ethical and legislative issues does AI face?
The main issue that exists today is that no one deeply thinks about the consequences of the race to Artifical General Intelligence, a human-like thinking machine that can address any issue and is fully aware of itself while constantly improving it’s abilities.
Because all firms are racing to dominate the world through this technology we cannot stop it. Since we never had any incident where human lives were lost as a result of such systems we have very few if at all attempts to legislate and involve governments in order to at least try and outline the safety guidelines of developing such systems.
One step forward is the treaty signed in Europe to ban autonomous weapons, which are weapons that move alone, think alone and decide which target to kill on its own without human intervention. However the US government didn’t sign or create such resolution of its own, nor did China or Russia.
This means that we might see a growing usage of such weapons. This is quite concerning. The ethical issues can come up on their own just by thinking about the consequences and the fact that there is no significant attempts to legislate the development of AGI.
Some opinion leaders say that our first incident with intelligent systems may be too late to legislate.
In what ways can AI and blockchain be integrated?
I think the best is to think outside buzzwords for anyone who is into new technologies. Blockchain is simply a transparent ledger system that can execute complex contracts between parties automatically.
Do AI systems need an infrastructure of Blockchain to perform any specific activity? Most likely not. It’s more the case where an existing blockchain project might want to analyse big data and teach an AI to perform certain tasks for their project.
There is of course the singularity network that basically wants to decentralise AI and create a space where developers can develop their own AIs and make them all communicate and update on one network. It’s a very ambitious project but it brings one example on how you can decentralise AI and make it belong to the community and not to some strong corporation like Google or IBM.
Is it better that way, or any safer? Only time will tell.
Does Malta have the necessary infrastructure to support AI efforts?
Malta is the fastest growing economy in Europe. I see great efforts to become AI- and blockchain-friendly. I would suggest that coming up with legislations of safety in AI can be a great step to show other countries that developing AI, especially AGI should not be done without keeping human interest in line.
What are Hyperion’s offerings regarding AI?
Hyperion can assist with setting up a team that can work on AI projects. Our base here in Malta provides solutions to companies that are looking to scale up on developers but can’t source them locally. Through our nearshore development centre in Ukraine we can assist and tackle any AI related project.
You will be at this May’s Malta A.I & Blockchain summit. What are your main objectives?
We are coming to meet our partners as well as form new connections with new partners that seek to innovate in fintech. I believe that the fintech industry in Malta lacks an edge and Hyperion can answer to that need.