Artificial Intelligence

AI, so hot right now

Artificial-intelligence-elon-musk-hawking

This blog post was originally posted at TechCrunch.

Artificial intelligence is one of the hottest subjects these days, and recent advances in technology makes AI even closer to reality than most of us can imagine. The subject really got traction when Stephen Hawking, Elon Musk and more than 1,000 AI and robotics researchers signed an open letter issuing a warning regarding the use of AI in weapons development last year. The following month BAE Systems unveiled Taranis, the most advanced autonomous UAV ever created, and there are currently 40 countries working on the deployment of AI in weapons development. The defense industry are not the only ones engaging in an arms race to create advanced AI.

Tech giants Facebook, Google, Microsoft and IBM are all engaging in various AI-initiatives as well as competing on developing digital personal assistants like Facebook’s M, Cortana from Microsoft and Apple’ Siri. Mark Zuckerberg even wants to create his own version of Jarvis from Iron Man to run his home. At this year’s World Economic Forum in Davos it was stated that artificial intelligence is ushering in the fourth industrial revolution that will change society as we know it and cost five million jobs by 2020.

Robots is no longer limited to traditional blue-collar jobs, fully automated assembly lines and high frequency trading algorithms. White-collar jobs are ripe for automation and robots are replacing bank tellers, mortgage brokers and loan officers in the financial industry. These examples follow strict repetitive rule-based routines and a machine easily performs that without any human interaction.

However, recent development is the beginning of a new era of AI that are able to perform complex tasks and no longer rely on pre-programmed rules in decision-making. Robo-advisor services like Betterment and Wealthfront are rising in popularity, and the hedge fund industry is launching AI-controlled funds that operates completely without human interaction. The co-head of one of these funds predicts that the time will come that no human investment manager will be able to beat the computer. But how is it possible for an AI to operate autonomously without any human interaction?

Machine learning is one of the fundamentals behind AI and was defined by Arthur Samuel back in 1959 as the science of getting computers to learn and act without being explicitly programmed. This technology is integral in the development of self-driving cars, IBM Watson, speech- and image recognition, as well as solving some of our most challenging tasks like making sense of the human genome. Machine learning has its roots from statistical pattern recognition, and is fundamental in many everyday applications and services like spam filters and web search algorithms. The fundamentals of machine learning is letting the computer program learn from examples. In order to accelerate machine learning development, Google released its machine learning system, Tensorflow on Github which led to Microsoft following up shortly after.

Deep learning takes the concept of machine learning even deeper (pun intended) and can model complex non-linear relationships consisting of many layers. Deep learning is often mentioned interchangeably together with artificial/deep neural networks, which can be viewed as biologically inspired programming paradigm, which enables a computer to learn from observational data. Deep learning is considered the techniques we apply to learn in neural networks.

Quantum computing is the latest and hottest in AI development and Google states that they have in collaboration with NASA a quantum Computer that is 100 million times faster than a traditional computer. The D-Wave 2X could theoretically complete calculations within seconds to a problem that might take a digital computer 10,000 years to calculate. Although Google states that quantum computing might not be suitable for deep learning. While traditional computers rely bits that are either 1 or 0, a quantum computer is based on qubits that can hold a superposition and be both 1 and 0 simultaneously. This state enables quantum computers to crunch data at an exponential rate. While quantum computing may not be suited for deep learning, it could revolutionize the field of optimization in logistics, investment strategies and energy production and consumption.

I have limited this post to include only a selection of the technologies and techniques applied in AI research and development. The road to artificial intelligence is no single discipline but a collection of specialized subject matters, techniques and theories that together interacts to create some form of intelligence. I have limited this post to include only a selection of the technologies and techniques applied in AI research and development. For further insight, I recommend looking into the online course offered by Google at Udacity on deep learning, the class on machine learning at Stanford Open Classroom as well as  evolutionary computing and logic programming (even though I still hold a grudge against Prolog for making me feel too stupid to really understand how it works).

Omfanget av roboter og kunstig intelligens er mer komplisert enn vi kan forestille oss

artificial-intelligence-698154_960_720

This blog post is in Norwegian and was originally published at Aftenposten.

Det var de beste av tider, det var de verste av tider. Det var visdommens tidsalder, det var dårskaps tidsalder. Digitalisering, delingsøkonomi og roboter var hovedtemaer under NHO sin årskonferanse og Norge rangeres av World Economic Forum som nummer 1 i verden når det kommer til digital infrastruktur.

Samtidig plasserer Harvard Business School Norge i kategorien for land som med høy digital modenhet, men med fare for å sakke akterut i evnen til å videreutvikle oss en digital økonomi, og Norge har stått stille mens verden har beveget seg videre.

Utviklingen av tingenes internett og smart bruk av robotteknologi bidrar til at stadig mer komplekse arbeidsoppgaver kan automatiseres, og delingsøkonomien fører til mer desentralisert produksjon og konsumpsjon.

Innen industrien er dette allerede en realitet hvor industrielle roboter lenge har dominert de fysiske samlebåndene, og dette vil i løpet av få år dominere de virtuelle samlebåndene knyttet til saksbehandling i både privat og offentlig sektor. Finansbransjen har allerede begynt å ta i bruk roboter til å utføre en rekke arbeidsoppgaver, og dette er ventet å omfang. Kunstig intelligens vil i økende grad bli en sentral del av både interne prosesser, produkter og tjenester, og det spås at roboter vil erstatte mennesker i en tredjedel av dagens tradisjonelle yrker innen 2025. Konsekvensen av dette er ifølge Jeremy Rifkin er slutten på kapitalismen slik vi kjenner den hvor roboter vil være i stand til å operere uten menneskelig interaksjon. Det første beviset på dette er Etherum Frontier Network som legger til rette for en desentralisert virksomhet gjennom bruk av blockchain-teknologien.

I en fremtid der roboter utfører selvstendige handlinger uten menneskelig interaksjon, hva skjer når en robot begår en kriminell handling? Hvem skal holdes ansvarlig for en slik handling? Dette kan høres ut som science fiction, men det har allerede skjedd.

Et sveitsisk kunstnerkollekvit opprettet en automatisert shopping robot med den hensikt å utføre en rekke tilfeldige kjøp på det mørke nettet. Roboten kjøpte en rekke varer, blant annet et ungarsk pass og en pose med ecstasy, før den ble “arrestert” av sveitsisk politi, uten at skaperne av roboten fikk noen konsekvenser.

Hvordan skal en robot reguleres når det opptrer på egen hånd uten menneskelig interaksjon? Marte Michelet tar til orde for behovet for en robotpolitikk, og det er allerede identifisert en rekke regulatoriske utfordringer knyttet til kontroll og regulering av kunstig intelligens. Hvis en robot skal holdes ansvarlig for sine handlinger, så vil vi måtte utstede en fysisk, juridisk og digital identitet på lik linje med et menneske. Hvis en robot gis det samme juridiske ansvaret som et menneske, bør ikke det har også de samme juridiske rettigheter som et menneske? Hvis en robot gis de samme rettighetene som et menneske, vil dette også inkludere retten til å reprodusere seg selv?

Det er sannsynlig at behovet for å finne svar på disse spørsmålene er nærmere enn vi tror. Google og IBM har startet et kappløp for å utvikle kunstig intelligens basert på kvantemaskiner, og i fjor sommer bestod en robot en selvbevissthetstest som tidligere var forbeholdt mennesker for første gang.

Kunstig intelligens og roboter reiser også en rekke etiske problemstillinger. Hvordan skal en selvkjørende bil håndtere en situasjon der den må velge mellom å kjøre på en fotgjenger eller sette bilens passasjer i fare ved å kjøre av veien?

Når vi snakker om roboter og kunstig intelligens er vi ofte begrenset til vår tradisjonelle oppfatning av verden hvor vi ser for oss roboter som menneskelignende androider. Dette er ikke nødvendigvis tilfelle; hvordan skal vi forholde oss til en selvbevisst kunstig intelligens som kun har en virtuell tilstedeværelse på et distribuert nettverk?

Kunstig intelligens reiser en rekke spørsmål på tvers av sosiale, økonomiske, politiske, teknologiske, juridiske, etiske og filosofiske problemstillinger. For å får en god nok forståelse av mulighetene og potensielle konsekvenser knyttet til roboter og kunstig intelligens er det behov for å forstå sammenhengene og konsekvensene på tvers av fagfelt.

Artificial intelligence, legal responsibility and civil rights

robot-507811_640

This post was originally published in TechCrunch

I have been a huge science fiction fan as long as I can remember, and a recurring theme in both science fiction literature and movies is the creation of artificial intelligence. However, the subject is becoming increasingly more science and less fiction with technological advancements. One of the earliest references to a robot, or an automaton is in The Iliad by Homer written sometime around 700 BC. With Isaac Asimov three laws of robotics, written all the way back in 1942. Arthur C. Clarke’s AI gone rogue in 2001: A Space Odyssey. The persecution of androids in Philip K. Dicks, Do android Dream of Electric Sheep (Better known as the movie Blade Runner). As well as depictions of AI in popular fiction through the iconic droids R2-D2 and C-3PO from Star Wars, Data from Star Trek and many more, the idea of creating artificial intelligence never seizes to amaze us.

This is a subject where my child like curiosity and interest exceeds my knowledge by a longshot, but recent development in AI and robotics both appeal to my inner child as well as my adult sense of skepticism. Along with the vast possibilities related to the creation of an artificial intelligence, there are also numerous challenges. In recent news, Stephen Hawking, Elon Musk and over 1,000 AI and robotics researchers signed an open letter issuing a warning regarding the use of AI in weapons development.

However, if we exclude the announced robot apocalypse for now, there are several other subjects to take into consideration.

Experts predict that robots will replace humans in one-third of today’s traditional professions by 2025, and according to Jeremy Rifkin, this will be the eclipse of capitalism as we know it. Machines will be self-replicating and able to operate as a hive mind without any human interaction, leading to a society where production is limited to the cost of raw materials.

What happens if an AI commit a crime? Who is responsible for the actions taken by an artificial intelligence? This may sound like science fiction, but has already happened. A Swiss art Group created an automated shopping robot with the purpose of committing random Darknet purchases. The robot managed to purchase several items, including a Hungarian passport and some Ecstasy pills before it was “arrested” by Swiss police. The aftermath resulted in no charges against neither the robot nor the artist behind the robot.

How should an AI be regulated when the AI is acting on its own outside the control of humans? There has already been identified several regulatory problems for controlling and regulating artificial intelligence. If an AI were to have legal responsibility for its actions, then an AI should have both a physical, legal and digital identity similar to a human being as well.

If an AI is given the same legal responsibilities as a human, shouldn’t it also have the same legal rights as a human? It is likely that an AI that has achieved self-awareness would demand equal rights instead of turning to killing all humans. If an AI were granted full civil rights, would that also include the right to reproduce? As a robot recently passed a self-awareness test previously believed that only humans could solve, the need to answer these questions may be closer than we think.

When addressing artificial intelligence, we are often limited to our traditional perception of the world, where we envision an AI as an android. This is not necessarily the case, how would we asses a self-aware and omnipresent digital intelligent being residing on a distributed computing network.

Artificial intelligence raises a series of questions across social, economic, political, technological, legal, ethical and philosophical issues. In order to untangle the uncertainties, possibilities and potential perils related to artificial intelligence, there is need to assess and understand the correlation between these fields. We should expect that, machines would continue to take us by surprise with great frequency.

Would you let a robot secure your financial future?

calculator-695084_1280

The financial services industry is facing massive changes in the years to come, and new technology is creating new possibilities every step of the way. Smartphones are changing the way we view payments and everyday banking, blockchain technology could revolutionize payment processing and marketplace lending and crowdfunding could redefine capital access for SMEs.

Artificial intelligence is another area that could become a central component in the future of financial services. While there is a lot of buzz surrounding AI, machine learning, cognitive computing and evolutionary algorithms, the use of algorithms is nothing new for the financial industry. Algorithmic trading has already made a huge impact on the stock markets, and credit and risk scoring algorithms has been at the core of banking for decades.

As algorithms are becoming more sophisticated, the potential applications are shifting from statistical analysis of historical data to a wide variety of potential applications.

The combination of data processing and natural language processing opens up new possibilities, and enables machines to see patterns in unstructured data, thus enabling banks to detect and prevent fraud and money laundering at an earlier stage.

The daily operation of today’s banks are composed of a set of automated processes with predefined conditions and rigid instructions, where bank tellers could in the utmost consequence be viewed as humans doing a robots job. For this machinery to stay compliant to an ever-increasing stack of governmental regulations case-workers and analysts are performing time-consuming and repetitive reporting processes, which also could benefit from the introduction of intelligent algorithms. The goal should be to not only automating the most mundane tasks, but allowing advanced algorithms to perform complex analysis and reporting. Many of these tasks are better suited for machines than humans anyway.

Lastly it is predicted that personal financial advisors will be replaced by robots in a not so distant future. This is where artificial intelligence will have the most visible impact on banking, and in this scenario, the future is now!

Regulatory compliance is challenging the business model of traditional financial advisors and wealth managers. Profitability is challenged by regulations like Retail Distribution Review (RDR), and the pressure on existing business models and need for automation will increase with the introduction of MiFID II.

One of the effects is what Deloitte refers to as an “advisory gap”, where only the wealthiest customers are profitable and eligible for traditional financial advice. This makes both the retail market and mass affluent segment up for grabs to whoever manages to offer user-friendly platforms combined with automated financial advice.

Enter the robo-advisor industry, which is expected to reach $2.2 trillion in assets under management by 2020, with a compound annual growth rate of 68%. The main driver behind this growth is considered to be the cost benefits of automating financial advice. There are several pros and cons when choosing between a robot and a human advisor, and some claim that robo-advisors are primarily suited for cost efficient processing of customers betting on low-cost index funds and ETFs.

This may be true from a short-term perspective, but I believe that the use of artificial intelligence in financial services should not be underestimated and viewed purely as a cost-saving measure.

The evolution in artificial intelligence is accelerating, and it could soon be difficult to distinguish between a human and robot advisor. For a first-hand experience, I recommend  trying out this Turing-test for investors.