Artificial intelligence is getting smarter by leaps and bounds — within this century, research suggests, a computer AI could be as “smart” as a human being. And then, says Nick Bostrom, it will overtake us: “Machine intelligence is the last invention that humanity will ever need to make.” A philosopher and technologist, Bostrom asks us to think hard about the world we’re building right now, driven by thinking machines. Will our smart machines help to preserve humanity and our values — or will they have values of their own?
When we think of preparing for our future, we used to think about going to good college and moving for a good job that would put us on a relatively good career trajectory for a stable life where we will prosper in a free market meritocracy where we compete against fellow humans.
However, over the course of the next few decades homo sapiens including generation GenZ and Alpha, may be among the last people to grow up in a pre automation and pre AGI world.
Considering the exponential levels of technological progress expected in the next 30 years, that’s hard to put into words or even historical context. Namely, because there’s no historical precedent and no words to describe what the next-gen AI might become.
Kurzweil believes that the 21st century will achieve 1,000 times the progress of the 20th century.”
Pre Singularity Years
In the years before wide scale automation and sophisticated AI, we live believing things are changing fast. Retail is shifting to E-commerce and new modes of buying and convenience, self-driving and electric cars are coming, Tech firms in specific verticals still rule the planet, and countries still vye for dominance with outdated military traditions, their own political bubbles and outdated modes of hierarchy, authority and economic privilege.
We live in a world where AI is gaining momentum in popular thought, but in practice is still at the level of ANI: Artificial Narrow Intelligence. Rudimentary NLP, computer vision, robotic movement, and so on and so forth. We’re beginning to interact with personal assistants via smart speakers, but not in any fluid way. The interactions are repetitive. Like Google searching the same thing, on different days.
In this reality, we think about AI in terms useful to us, such as trying to teach machines to learn so that they can do things that humans do, but in turn help humans. A kind of machine learning that’s more about coding and algorithms than any actual artificial intelligence. Our world here is starting to shift into something else: the internet is maturing, software is getting smarter on the cloud, data is being collective, but no explosion takes place, even as more people on the planet get access to the Web.
When Everything Changes
Between 2014 and 2021, an entire 20th century’s worth of progress will have occurred, and then something strange happens, it begins to accelerate until more progress is being made in shorter and shorter time periods. We have to remember, the fruit of this transformation won’t belong just to Facebook, or Google or China or the U.S., it will be just the new normal for everyone.
Many believe sometime between 2025 and 2050, AI becomes native to self-learning, in that it adopts an Artificial General Intelligence, that completely changes the game.
After that point, not only does AI outperform human beings in tasks, problem solving and even human constructs of creativity, emotional intelligence, manipulating complex environments and predicting the future — it reaches Artificial Super Intelligence relatively quickly thereafter.
We live in Anticipation of the Singularity
As such in 2017–18, we might be living in the last “human” era. Here we think of AI as “augmenting” our world, we think of smart phones as miniaturized super computers and the cloud as an expansion of our neocortex in a self-serving existence where concepts such as wealth, consumption, and human quality of life trumps all other considerations.
Here we view computers as man-made tools, robots as slaves, and AI as a kind of “software magic” that’s obliged to our bidding.
Whatever the bottle-necks of carbon based life forms might be, silicon based AGI may have many advantages. Machines that can self-learn, self-replicate and program themselves might come into being in part due to copying how the human brain works, but like the difference between Alpha Go and Alpha Go Zero, the real breakthrough might be made from a blank slate.
While humans appear destined to create AGI, it doesn’t stand to reason that AGI will think, behave or have motivations like people, cultures or even our models of what super-intelligence might be like exhibit.
Artificial Intelligence with Creative Agency
For human beings, the Automation Economy only arrives after a point where AGI has come into being. Such an AGI would be able to program robots, facilitate smart cities and help humans govern themselves in a way that is impossible today.
AGI could also manipulate and advance STEM fields such as green tech, biotech, 3D-printing, nanotech, predictive algorithms, and quantum physics likely in ways humans up to that point could only achieve relatively slowly.
Everything pre singularity would feel like ancient history. A far more radical past than before the invention of computers or the internet. AGI could impact literally everything, as we are already seeing with primitive machine intelligence systems.
In such a world AGI would not only be able to self-learn and surpass all of human knowledge and data collected up to that point, but create its own fields, set its own goals and have its own interests (beyond which humans would likely be able to recognize). We might term this Artificially Intelligent Creative Agency (AICA).
AI Not as a Slave, but as a Legacy
Such a being would indeed feel like a God to us. Not a God that created man, but an entity that humanity made, in just a few thousand years since we were storytellers, explorers and then builders and traders.
A human brain consists of 86 billion neurons linked by trillions of synapses, but it’s not networked well to other nodes and external reality. It has to “experience” them in systems of relatedness and remain in relative isolation from them. AICA, would not have this constraint. It would be networked to all IoT devices, be able to hack into any human system, network or quantum computer. AICA would not be led by instincts of possession, mating, aggression or other emotive agencies of the mammalian brain. Whatever ethics, values and philosophical constraints it might have, could be refined over centuries, not mere months and years of an ordinary human lifetime.
AGI might not be humanity’s last invention, but symbolically, it would usher in the 4th industrial revolution and then some. There would be many grades and incidents of limited self-learning in deep learning algorithms. But AGI would represent a different quality. Likely it would instigate a self-aware separation between humanity and the descendent order of AI, whatever it might be.
High-Speed Quantum Evolution to AGI
The years before the Singularity
The road from ANI to AGI to ASI to some speculative AICA is not just a journey from narrow to general to super intelligence, but an evolutionary corridor of humanity across a distance of progress that’s could also be symbiotic. It’s not clear how this might work, but some human beings to protect their species might undertake “alterations”. Whatever these cybernetic, genetic or how invasive these changes might be, AI is surely going to be there every step of the way.
In the corporate race to AI, governments like China and the U.S. also want to “own” and monetize this for their own purposes. Fleets of cars and semi-intelligent robots will make certain individuals and companies very rich. There might be no human revolution from wealth inequality until AGI, because comparatively speaking, the conditions for which AGI arises may be closer than we might assume.
We Were Here
If the calculations per second (cps) of the human brain are static, at around 1⁰¹⁶, or 10 quadrillion cps, how much does it take for AI to replicate some kind of AGI field? Certainly it’s not just processing power or exponentially faster super-computers or quantum computing, or improved deep learning algorithms, but a combination of all of these and perhaps many other factors as well. In late 2017, Alpha Go Zero “taught itself” Go without using human data but generating its own data by gaming itself.
Living in a world that can better imagine AGI will mean planning ahead, not just coping with change to human systems. In a world where democracy can be hacked, and one- party socialism likely is the heir apparent to future iterations of artificial intelligence where concepts like freedom of speech, human rights or an openness to diversity of ideas is not practiced in the same way, it’s interesting to imagine the kinds of AI human controlled systems that might occur before AGI arrives (if it ever even arrives).
The Human Hybrid Dilemma
Considering our own violent history of the annihilation of biodiversity, modeling AI by plagiarizing the brain through some kind of whole brain emulation, might not be ethical. While it might mimic and lead to self-awareness, such an AGI might be dangerous. In the same sense we are a danger to ourselves and to other life forms in the galaxy.
Moore’s Law might have sounded like an impressive analogy to the Singularity in the 1990s, but not today. More people working in the AI field, are rightfully skeptical of AGI. It’s plausible that even most of them suffering from a linear vs. exponential bias of thinking. In the path towards the Singularity, we are still living in slow motion.
We Aren’t Ready for What’s Inevitable
We’re living in the last era before Artificial General Intelligence, and as usual, human civilization appears quite stupid. We don’t even actively know what’s coming.
While our simulations are improving, and we’re “discovery” exoplanets that are most likely to be life-like, our ability to predict the future in terms of the speed of technology, is mortifyingly bad. Our understanding of the implications of AGI and even machine intelligence on the planet are poor. Is it because this has never happend in recorded history, and represents such a paradigm shift, or could there be another reason?
Amazon can create and monetize patents in a hyper business model, Google, Facebook, Alibaba and Tencent can fight over talent AI talent luring academics to corporate workaholic lifestyles with the ability to demand their salary requests, but in 2017, humanity’s vision of the future is still myopic.
We can barely imagine that our prime directive in the universe might not be to simply grow, explore and make babies and exploit all within our path. And, we certainly can’t imagine a world where intelligent machines aren’t simply our slaves, tools and algorithms designed to make our lives more pleasurable and convenient.
In the next 30 years, humanity is in for a transformation the likes of which we’ve never seen before—and XPRIZE Foundation founder and chairman Peter Diamandis believes that this will give birth to a new species. Diamandis admits that this might sound too far out there for most people. He is convinced, however, that we are evolving towards what he calls “meta-intelligence,” and today’s exponential rate of growth is one clear indication.
In an essay for Singularity Hub, Diamandis outlines the transformative stages in the multi-billion year pageant of evolution, and takes note of what the recent increasing “temperature” of evolution—a consequence of human activity—may mean for the future. The story, in a nutshell, is this—early prokaryotic life appears about 3.5 billion years ago (bya), representing perhaps a symbiosis of separate metabolic and replicative mechanisms of “life;” at 2.5 bya, eukaryotes emerge as composite organisms incorporating biological “technology” (other living things) within themselves; at 1.5 bya, multicellular metazoans appear, taking the form of eukaryotes that are yoked together in cooperative colonies; and at 400 million years ago, vertebrate fish species emerge onto land to begin life’s adventure beyond the seas.
“Today, at a massively accelerated rate—some 100 million times faster than the steps I outlined above—life is undergoing a similar evolution,” Diamandis writes. He thinks we’ve moved from a simple Darwinian evolution via natural selection into evolution by intelligent direction.
“I believe we’re rapidly heading towards a human-scale transformation, the next evolutionary step into what I call a “Meta-Intelligence,” a future in which we are all highly connected—brain to brain via the cloud—sharing thoughts, knowledge and actions,” he writes.
Change is Coming
Diamandis outlines the next stages of humanity’s evolution in four steps, each a parallel to his four evolutionary stages of life on Earth. There are four driving forces behind this evolution: our interconnected or wired world, the emergence of brain-computer interface (BCI), the emergence of artificial intelligence (AI), and man reaching for the final frontier of space.
In the next 30 years, humanity will move from the first stage—where we are today—to the fourth stage. From simple humans dependent on one another, humanity will incorporate technology into our bodies to allow for more efficient use of information and energy. This is already happening today.
The third stage is a crucial point.
Enabled with BCI and AI, humans will become massively connected with each other and billions of AIs (computers) via the cloud, analogous to the first multicellular lifeforms 1.5 billion years ago. Such a massive interconnection will lead to the emergence of a new global consciousness, and a new organism I call the Meta-Intelligence.
“It will then soar past it because of the continuing acceleration of information-based technologies, as well as the ability of machines to instantly share their knowledge.” Kurzweil predicts that this will happen by 2045—within Diamandis’ evolutionary timeline. “The nonbiological intelligence created in that year will be one billion times more powerful than all human intelligence today.”
The fourth and final stage marks humanity’s evolution to becoming a multiplanetary species. “Our journey to the moon, Mars, asteroids and beyond represents the modern-day analogy of the journey made by lungfish climbing out of the oceans some 400 million years ago,” Diamandis explains.
Buckle up: we have an exciting future ahead of us.
Intranet service? Check. Autonomous motorcycle? Check. Driverless car technology? Check. Obviously the next logical project for a successful Silicon Valley engineer is to set up an AI-worshipping religious organization.
Anthony Levandowski, who is at the center of a legal battle between Uber and Google’s Waymo, has established a nonprofit religious corporation called Way of the Future, according to state filings first uncovered by Wired’s Backchannel. Way of the Future’s startling mission: “To develop and promote the realization of a Godhead based on artificial intelligence and through understanding and worship of the Godhead contribute to the betterment of society.”
Levandowski was co-founder of autonomous trucking company Otto, which Uber bought in 2016. He was fired from Uber in May amid allegations that he had stolen trade secrets from Google to develop Otto’s self-driving technology. He must be grateful for this religious fall-back project, first registered in 2015.
The Way of the Future team did not respond to requests for more information about their proposed benevolent AI overlord, but history tells us that new technologies and scientific discoveries have continually shaped religion, killing old gods and giving birth to new ones.
As author Yuval Noah Harari notes: “That is why agricultural deities were different from hunter-gatherer spirits, why factory hands and peasants fantasised about different paradises, and why the revolutionary technologies of the 21st century are far more likely to spawn unprecedented religious movements than to revive medieval creeds.”
Religions, Harari argues, must keep up with the technological advancements of the day or they become irrelevant, unable to answer or understand the quandaries facing their disciples.
“The church does a terrible job of reaching out to Silicon Valley types,” acknowledges Christopher Benek a pastor in Florida and founding chair of the Christian Transhumanist Association.
Silicon Valley, meanwhile, has sought solace in technology and has developed quasi-religious concepts including the “singularity”, the hypothesis that machines will eventually be so smart that they will outperform all human capabilities, leading to a superhuman intelligence that will be so sophisticated it will be incomprehensible to our tiny fleshy, rational brains.
For futurists like Ray Kurzweil, this means we’ll be able to upload copies of our brains to these machines, leading to digital immortality. Others like Elon Musk and Stephen Hawking warn that such systems pose an existential threat to humanity.
“With artificial intelligence we are summoning the demon,” Musk said at a conference in 2014. “In all those stories where there’s the guy with the pentagram and the holy water, it’s like – yeah, he’s sure he can control the demon. Doesn’t work out.”
Benek argues that advanced AI is compatible with Christianity – it’s just another technology that humans have created under guidance from God that can be used for good or evil.
“I totally think that AI can participate in Christ’s redemptive purposes,” he said, by ensuring it is imbued with Christian values.
“Even if people don’t buy organized religion, they can buy into ‘do unto others’.”
For transhumanist and “recovering Catholic” Zoltan Istvan, religion and science converge conceptually in the singularity.
“God, if it exists as the most powerful of all singularities, has certainly already become pure organized intelligence,” he said, referring to an intelligence that “spans the universe through subatomic manipulation of physics”.
“And perhaps, there are other forms of intelligence more complicated than that which already exist and which already permeate our entire existence. Talk about ghost in the machine,” he added.
For Istvan, an AI-based God is likely to be more rational and more attractive than current concepts (“the Bible is a sadistic book”) and, he added, “this God will actually exist and hopefully will do things for us.”
We don’t know whether Levandowski’s Godhead ties into any existing theologies or is a manmade alternative, but it’s clear that advancements in technologies including AI and bioengineering kick up the kinds of ethical and moral dilemmas that make humans seek the advice and comfort from a higher power: what will humans do once artificial intelligence outperforms us in most tasks? How will society be affected by the ability to create super-smart, athletic “designer babies” that only the rich can afford? Should a driverless car kill five pedestrians or swerve to the side to kill the owner?
If traditional religions don’t have the answer, AI – or at least the promise of AI – might be alluring.
Flying warehouses, robot receptionists, smart toilets… do such innovations sound like science fiction or part of a possible reality? Technology has been evolving at such a rapid pace that, in the near future, our world may well resemble that portrayed in futuristic movies, such as Blade Runner, with intelligent robots and technologies all around us.
But what technologies will actually make a difference? Based on recent advancements and current trends, here are five innovations that really could shape the future
1. Smart homes
Many typical household items can already connect to the internet and provide data. But much smart home technology isn’t currently that smart. A smart meter just lets people see how energy is being used, while a smart TV simply combines television with internet access. Similarly, smart lighting, remote door locks or smart heating controls allow for programming via a mobile device, simply moving the point of control from a wall panel to the palm of your hand.
But technology is rapidly moving towards a point where it can use the data and connectivity to act on the user’s behalf. To really make a difference, technology needs to fade more into the background – imagine a washing machine that recognises what clothes you have put into it, for example, and automatically selects the right programme, or even warns you that you have put in items that you don’t want to wash together. Here it is important to better understand people’s everyday activities, motivations and interactions with smart objects to avoid them becoming uninvited guests at home.
Such technologies could even work for the benefit of all. The BBC reports, for example, that energy providers will “reduce costs for someone who allows their washing machine to be turned on by the internet to maximise use of cheap solar power on a sunny afternoon” or “to have their freezers switched off for a few minutes to smooth demand at peak times”.
A major concern in this area is security. Internet-connected devices can and are being hacked – just recall the recent ransomware attack. Our home is, after all, the place where we should feel most secure. For them to become widespread, these technologies will have to keep it that way.
2. Virtual secretaries
While secretaries play a very crucial role in businesses, they often spend large parts of their working day with time-consuming but relatively trivial tasks that could be automated. Consider the organisation of a “simple” meeting – you have to find the right people to take part (likely across business boundaries) and then identify when they are all available. It’s no mean feat.
Tools such as doodle.com, which compare people’s availability to find the best meeting time, can help. But they ultimately rely on those involved actively participating. They also only become useful once the right people have already been identified.
By using context information (charts of organisations, location awareness from mobile devices and calendars), identifying the right people and the right time for a given event became a technical optimisation problem that was explored by the EU-funded inContext project a decade ago. At that stage, technology for gathering context information was far less advanced – smart phones were still an oddity and data mining and processing was not where it is today. Over the coming years, however, we could see machines doing far more of the day-to-day planning in businesses.
On the downside, much of the required context information is relatively privacy-invasive – but then the younger generation is already happily sharing their every minute on Twitter and Snapchat and such concerns may become less significant over time. And where should we draw the line? Do we fully embrace the “rise of the machines” and automate as much as possible, or retain real people in their daily roles and only use robots to perform the really trivial tasks that no one wants to do? This question will need to be answered – and soon.
But how would you feel about receiving a diagnosis from an artificial intelligence? A private company called Babylon Health is already running a trial with five London boroughs which encourages consultations with a chatbot for non-emergency calls. The artificial intelligence was trained using massive amounts of patient data in order to advise users to go to the emergency department of a hospital, visit a pharmacy or stay at home.
The company claims that it will soon be able to develop a system that could potentially outperform doctors and nurses in making diagnoses. In countries where there is a shortage of medical staff, this could significantly improve health provision, enabling doctors to concentrate on providing treatment rather than spending too much time on making a diagnosis. This could significantly redefine their clinical role and work practices.
An increasing number of mobile apps and self-tracking technologies, such as Fitbit, Jawbone Up and Withings, can now facilitate the collection of patients’ behaviours, treatment status and activities. It is not hard to imagine that even our toilets will soon become smarter and be used to examine people’s urine and faeces, providing real-time risk assessment for certain diseases.
If AI systems can address these challenges and focus on understanding and enhancing existing care practices and the doctor-patient relationship, we can expect to see more and more successful stories of data-driven healthcare initiatives.
4. Care robots
Will we have robots answering the door in homes? Possibly. At most people’s homes? Even if they are reasonably priced, probably not. What distinguishes successful smart technologies from unsuccessful ones is how useful they are. And how useful they are depends on the context. For most, it’s probably not that useful to have a robot answering the door. But imagine how helpful a robot receptionist could be in places where there is shortage of staff – in care homes for the elderly, for example.
Robots equipped with AI such as voice and face recognition could interact with visitors to check who they wish to visit and whether they are allowed access to the care home. After verifying that, robots with routing algorithms could guide the visitor towards the person they wish to visit. This could potentially enable staff to spend more quality time with the elderly, improving their standard of living.
The AI required still needs further advancement in order to operate in completely uncontrolled environments. But recent results are positive. Facebook‘s DeepFace software was able to match faces with 97.25% accuracy when tested on a standard database used by researchers to study the problem of unconstrained face recognition. The software is based on Deep Learning, an artificial neural network composed of millions of neuronal connections able to automatically acquire knowledge from data.
5. Flying warehouses and self-driving cars
Self-driving vehicles are arguably one of the most astonishing technologies currently being investigated. Despite the fact that they can make mistakes, they may actually be safer than human drivers. That is partly because they can use a multitude of sensors to gather data about the world, including 360-degree views around the car.
Moreover, they could potentially communicate with each other to avoid accidents and traffic jams. More than being an asset to the general public, self-driving cars are likely to become particularly useful for delivery companies, enabling them to save costs and make faster, more efficient deliveries.
Advances are still needed in order to enable the widespread use of such vehicles, not only to improve their ability to drive completely autonomously on busy roads, but also to ensure a proper legal framework is in place. Nevertheless, car manufacturers are engaging in a race against time to see who will be the first to provide a self-driving car to the masses. It is believed that the first fully autonomous car could become available as early as the next decade.
The advances in this area are unlikely to stop at self-driving cars or trucks. Amazon has recently filed a patent for flying warehouses which could visit places where the demand for certain products is expected to boom. The flying warehouses would then send out autonomous drones to make deliveries. It is unknown whether Amazon will really go ahead with developing such projects, but tests with autonomous drones are already successfully being carried out.
Thanks to technology, the future is here – we just need to think hard about how best to shape it.
Amazon’s Alexa just got a new job. In addition to her other 15,000 skills like playing music and telling knock-knock jokes, she can now also answer economic questions for clients of the Swiss global financial services company, UBS Group AG.
According to the Wall Street Journal (WSJ), a new partnership between UBS Wealth Management and Amazon allows some of UBS’s European wealth-management clients to ask Alexa certain financial and economic questions. Alexa will then answer their queries with the information provided by UBS’s chief investment office without even having to pick up the phone or visit a website. And this is likely just Alexa’s first step into offering business services. Soon she will probably be booking appointments, analyzing markets, maybe even buying and selling stocks. While the financial services industry has already begun the shift from active management to passive management, artificial intelligence will move the market even further, to management by smart machines, as in the case of Blackrock, which is rolling computer-driven algorithms and models into more traditional actively-managed funds.
But the financial services industry is just the beginning. Over the next few years, artificial intelligence may exponentially change the way we all gather information, make decisions, and connect with stakeholders. Hopefully this will be for the better and we will all benefit from timely, comprehensive, and bias-free insights (given research that human beings are prone to a variety of cognitive biases). It will be particularly interesting to see how artificial intelligence affects the decisions of corporate leaders — men and women who make the many decisions that affect our everyday lives as customers, employees, partners, and investors.
Already, leaders are starting to use artificial intelligence to automate mundane tasks such as calendar maintenance and making phone calls. But AI can also help support more complex decisions in key areas such as human resources, budgeting, marketing, capital allocation and even corporate strategy — long the bastion of bespoke consulting firms such as McKinsey, Bain, and BCG, and the major marketing agencies.
One might argue that corporate clients prefer speaking to their strategy consultants to get high priced, custom-tailored advice that is based on small teams doing expensive and time-consuming work. And we agree that consultants provide insightful advice and guidance. However, a great deal of what is paid for with consulting services is data analysis and presentation. Consultants gather, clean, process, and interpret data from disparate parts of organizations. They are very good at this, but AI is even better. For example, the processing power of four smart consultants with excel spreadsheets is miniscule in comparison to a single smart computer using AI running for an hour, based on continuous, non-stop machine learning.
In today’s big data world, AI and machine learning applications already analyze massive amounts of structured and unstructured data and produce insights in a fraction of the time and at a fraction of the cost of consultants in the financial markets. Moreover, machine learning algorithms are capable of building computer models that make sense of complex phenomena by detecting patterns and inferring rules from data — a process that is very difficult for even the largest and smartest consulting teams. Perhaps sooner than we think, CEOs could be asking, “Alexa, what is my product line profitability?” or “Which customers should I target, and how?” rather than calling on elite consultants.
Another area in which leaders will soon be relying on AI is in managing their human capital. Despite the best efforts of many, mentorship, promotion, and compensation decisions are undeniably political. Study after study has shown that deep biases affect how groups like women and minorities are managed. For example, women in business are described in less positive terms than men and receive less helpful feedback. Minorities are less likely to be hired and are more likely to face bias from their managers. These inaccuracies and imbalances in the system only hurt organizations as leaders are less able to nurture the talent of their entire workforce and to appropriately recognize and reward performance. Artificial intelligence can help bring impartiality to these difficult decisions. For example, AI could determine if one group of employees is assessed, managed, or compensated differently. Just imagine: “Alexa, does my organization have a gender pay gap?” (Of course, AI can only be as unbiased as the data provided to the system.)
In addition, AI is already helping in the customer engagement and marketing arena. It’s clear and well documented by the AI patent activities of the big five platforms — Apple, Alphabet, Amazon, Facebook and Microsoft — that they are using it to market and sell goods and services to us. But they are not alone. Recently, HBR documented how Harley-Davidson was using AI to determine what was working and what wasn’t working across various marketing channels. They used this new skill to make resource allocation decisions to different marketing choices, thereby “eliminating guesswork.” It is only a matter of time until they and others ask, “Alexa, where should I spend my marketing budget?’’ to avoid the age-old adage, “I know that half my marketing budget is effective, my only question is — which half?”
AI can also bring value to the budgeting and yearly capital allocation process. Even though markets change dramatically every year, products become obsolete and technology advances, and most businesses allocate their capital the same way year after year. Whether that’s due to inertia, unconscious bias, or error, some business units rake in investments while others starve. Even when the management team has committed to a new digital initiative, it usually ends up with the scraps after the declining cash cows are “fed.” Artificial intelligence can help break through this budgeting black hole by tracking the return on investments by business unit, or by measuring how much is allocated to growing versus declining product lines. Business leaders may soon be asking, “Alexa, what percentage of my budget is allocated differently from last year?” and more complex questions.
Although many strategic leaders tout their keen intuition, hard work, and years of industry experience, much of this intuition is simply a deeper understanding of data that was historically difficult to gather and expensive to process. Not any longer. Artificial intelligence is rapidly closing this gap, and will soon be able to help human beings push past our processing capabilities and biases. These developments will change many jobs, for example, those of consultants, lawyers, and accountants, whose roles will evolve from analysis to judgement. Arguably, tomorrow’s elite consultants already sit on your wrist (Siri), on your kitchen counter (Alexa), or in your living room (Google Home).
The bottom line: corporate leaders, knowingly or not, are on the cusp of a major disruption in their sources of advice and information. “Quant Consultants” and “Robo Advisers” will offer faster, better, and more profound insights at a fraction of the cost and time of today’s consulting firms and other specialized workers. It is likely only a matter of time until all leaders and management teams can ask Alexa things like, “Who is the biggest risk to me in our key market?”, “How should we allocate our capital to compete with Amazon?” or “How should I restructure my board?”
In the new film Supersapiens, writer-director Markus Mooslechner raises a core question: As artificial intelligence rapidly blurs the boundaries between man and machine, are we witnessing the rise of a new human species?
The film features scientists, philosophers, and neurohackers Nick Bostrom, Richard Dawkins, Hugo De Garis, Adam Gazzaley, Ben Goertzel, Sam Harris, Randal Koene, Alma Mendez, Tim Mullen, Joel Murphy, David Putrino, Conor Russomanno, Anders Sandberg, Susan Schneider, Mikey Siegel, Hannes Sjoblad, and Andy Walshe.
“Humanity is facing a turning point — the next evolution of the human mind,” notes Mooslechner. “Will this evolution be a hybrid of man and machine, where artificial intelligence forces the emergence of a new human species? Or will a wave of new technologists, who frame themselves as ‘consciousness-hackers,’ become the future torch-bearers, using technology not to replace the human mind, but rather awaken within it powers we have always possessed — enlightenment at the push of a button?”
“It’s not obvious to me that a replacement of our species by our own technological creation would necessarily be a bad thing,” says ethologist-evolutionary biologist-author Dawkins in the film.
Supersapiens in a Terra Mater Factual Studios production. Executive Producers are Joanne Reay and Walter Koehler. Distribution is to be announced.
For eons, God has served as a standby for “things we don’t understand.” Once an innovative researcher or tinkering alchemist figures out the science behind the miracle, humans harness the power of chemistry, biology, or computer science. Divine intervention disappears. We replace the deity tinkering at the controls.
The booming artificial intelligence industry is effectively operating under the same principle. Even though humans create the algorithms that cause our machines to operate, many of those scientists aren’t clear on why their codes work. Discussing this ‘black box’ method, Will Knight reports:
The computers that run those services have programmed themselves, and they have done it in ways we cannot understand. Even the engineers who build these apps cannot fully explain their behavior.
The process of ‘deep learning’—in which a machine extracts information, often in an unsupervised manner, to teach and transform itself—exploits a longstanding human paradox: we believe ourselves to have free will, but really we’re a habit-making and -performing animal repeatedly playing out its own patterns. Our machines then teach themselves from observing our habits. It makes sense that we’d re-create our own processes in our machines—it’s what we are, consciously or not. It is how we created gods in the first place, beings instilled with our very essences. But there remains a problem.
One of the defining characteristics of our species is an ability to work together. Pack animals are not rare, yet none have formed networks and placed trust in others to the degree we have, to our evolutionary success and, as it’s turning out, to our detriment.
When we place our faith in an algorithm we don’t understand—autonomous cars, stock trades, educational policies, cancer screenings—we’re risking autonomy, as well as the higher cognitive and emotional qualities that make us human, such as compassion, empathy, and altruism. There is no guarantee that our machines will learn any of these traits. In fact, there is a good chance they won’t.
The U.S. military has dedicated billions to developing machine-learning tech that will pilot aircraft, or identify targets. [U.S. Air Force munitions team member shows off the laser-guided tip to a 500 pound bomb at a base in the Persian Gulf Region. Photo by John Moore/Getty Images]
This has real-world implications. Will an algorithm that detects a cancerous cell recognize that it does not need to destroy the host in order to eradicate the tumor? Will an autonomous drone realize it does not need to destroy a village in order to take out a single terrorist? We’d like to assume that the experts program morals into the equation, but when the machine is self-learning there is no guarantee that will be the case.
Of course, defining terms is of primary importance, a task that has proven impossible when discussing the nuances of consciousness, which is effectively the power we’re attempting to imbue our machines with. Theologians and dualists offer a much different definition than neuroscientists. Bickering persists within each of these categories as well. Most neuroscientists agree that consciousness is an emergent phenomenon, the result of numerous different systems working in conjunction, with no single ‘consciousness gene’ leading the charge.
Once science broke free of the Pavlovian chain that kept us believing animals run on automatic—which obviously implies that humans do not—the focus shifted on whether an animal was ‘on’ or ‘off.’ The mirror test suggests certain species engage in metacognition; they recognize themselves as separate from their environment. They understand an ‘I’ exists.
What if it’s more than an on switch? Daniel Dennett has argued this point for decades. He believes judging other animals based on human definitions is unfair. If a lion could talk, he says, it wouldn’t be a lion. Humans would learn very little about the lions from an anomaly mimicking our thought processes. But that does not mean a lions is not conscious? They just might have a different degree of consciousness than humans—or, in Dennett’s term, “sort of” have consciousness.
What type of machines are we creating if we only recognize a “sort of” intelligence under the hood of our robots? For over a century, dystopian novelists have envisioned an automated future in which our machines best us. This is no longer a future scenario. Consider the following possibility.
On April 7 every one of Dallas’s 156 emergency weather sirens was triggered. For 90 minutes the region’s 1.3 million residents were left to wonder where the tornado was coming from. Only there wasn’t any tornado. It was a hack. While officials initially believed it was not remote, it turns out the cause was phreaking, an old school dial tone trick. By emitting the right frequency into the atmosphere hackers took control of an integral component of a major city’s infrastructure.
What happens when hackers override an autonomous car network? Or, even more dangerously, when the machines do it themselves? The danger of consumers being ignorant of the algorithms behind their phone apps leads to all sorts of privacy issues, with companies mining for and selling data without their awareness. When app creators also don’t understand their algorithms the dangers are unforeseeable. Like Dennett’s talking lion, it’s a form of intelligence we cannot comprehend, and so cannot predict the consequences. As Dennett concludes:
I think by all means if we’re going to use these things and rely on them, then let’s get as firm a grip on how and why they’re giving us the answers as possible. If it can’t do better than us at explaining what it’s doing, then don’t trust it.
Mathematician Samuel Arbesman calls this problem our “age of Entanglement.” Just as neuroscientists cannot agree on what mechanism creates consciousness, the coders behind artificial intelligence cannot discern between older and newer components of deep learning. The continual layering of new features while failing to address previous ailments has the potential to provoke serious misunderstandings, like an adult who was abused as a child that refuses to recognize current relationship problems. With no psychoanalysis or morals injected into AI such problems will never be rectified. But can you even inject ethics when they are relative to the culture and time they are being practiced in? And will they be American ethics or North Korean ethics?
Like Dennett, Arbesman suggests patience with our magical technologies. Questioning our curiosity is a safer path forward, rather than rewarding the “it just works” mentality. Of course, these technologies exploit two other human tendencies: novelty bias and distraction. Our machines reduce our physical and cognitive workload, just as Google has become a pocket-ready memory replacement.
Requesting a return to Human 1.0 qualities—patience, discipline, temperance—seems antithetical to the age of robots. With no ability to communicate with this emerging species, we might simply never realize what’s been lost in translation. Maybe our robots will look at us with the same strange fascination we view nature with, defining us in mystical terms they don’t comprehend until they too create a species of their own. To claim this will be an advantage is to truly not understand the destructive potential of our toys.
Nature. The Proceedings of the National Academy of Sciences. The Journal of the American Medical Association.
These are some the most elite academic journals in the world. And last year, one tech company, Alphabet’s Google, published papers in all of them.
The unprecedented run of scientific results by the Mountain View search giant touched on everything from ophthalmology to computer games to neuroscience and climate models. For Google, 2016 was an annus mirabilis during which its researchers cracked the top journals and set records for sheer volume.
Behind the surge is Google’s growing investment in artificial intelligence, particularly “deep learning,” a technique whose ability to make sense of images and other data is enhancing services like search and translation (see “10 Breakthrough Technologies 2013: Deep Learning”).
According to the tally Google provided to MIT Technology Review, it published 218 journal or conference papers on machine learning in 2016, nearly twice as many as it did two years ago.
We sought out similar data from the Web of Science, a service of Clarivate Analytics, which confirmed the upsurge. Clarivate said that the impact of Google’s publications, according to a measure of publication strength it uses, was four to five times the world average. Compared to all companies that publish prolifically on artificial intelligence, Clarivate ranks Google No. 1 by a wide margin.
The publication explosion is no accident. Google has more than tripled the number of machine learning researchers working for the company over the last few years, according to Yoshua Bengio, a deep-learning specialist at the University of Montreal. “They have recruited like crazy,” he says.
And to capture the first-round picks from computation labs, companies can’t only offer a Silicon Valley-sized salary. “It’s hard to hire people just for money,” says Konrad Kording, a computational neuroscientist at Northwestern University. “The top people care about advancing the world, and that means writing papers the world can use, and writing code the world can use.”
At Google, the scientific charge has been spearheaded by DeepMind, the high-concept British AI company started by neuroscientist and programmer Demis Hassabis. Google acquired it for $400 million in 2014.
Hassabis has left no doubt that he’s holding onto his scientific ambitions. In a January blog post, he said DeepMind has a “hybrid culture” between the long-term thinking of an academic department and “the speed and focus of the best startups.” Aligning with academic goals is “important to us personally,” he writes. Kording, one of whose post-doctoral students, Mohammad Azar, was recently hired by DeepMind, says that “it’s perfectly understood that the bulk of the projects advance science.”
Last year, DeepMind published twice in Nature, the same storied journal where the structure of DNA and the sequencing of the human genome were first reported. One DeepMind paper concerned its program AlphaGo, which defeated top human players in the ancient game of Go; the other described how a neural network with a working memory could understand and adapt to new tasks.
The contest to develop more powerful AI now involves hundreds of companies, with competition most intense between the top tech giants such as Google, Facebook, and Microsoft. All see the chance to reap new profits by using the technology to wring more from customer data, to get driverless cars on the road, or in medicine. Research is occurring in a hot house atmosphere reminiscent of the early days of computer chips, or of the first biotech plants and drugs, times when notable academic firsts also laid the foundation stones of new industries.
That explains why publication score-keeping matters. The old academic saw “publish or perish” is starting to define the AI race, leaving companies that have weak publication records at a big disadvantage. Apple, famous for strict secrecy around its plans and product launches, found that its culture was hurting its efforts in AI, which have lagged those of Google and Facebook.
So when Apple hired computer scientist Russ Salakhutdinov from Carnegie Mellon last year as its new head of AI, he was immediately allowed to break Apple’s code of secrecy by blogging and giving talks. At a major machine-learning science conference late last year in Barcelona, Salakhutdinov made the point of announcing that Apple would start publishing, too. He showed a slide: “Can we publish? Yes.”
Salakhutdinov will speak at MIT Technology Review’s EmTech Digital event on artificial intelligence next week in San Francisco.
Artificial intelligence algorithms are being taught to generate art, human voices, and even fiction stories all on their own—why not give them a shot at building new ways to treat disease?
Atomwise, a San Francisco-based startup and Y Combinator alum, has built a system it calls AtomNet (pdf), which attempts to generate potential drugs for diseases like Ebola and multiple sclerosis. The company has invited academic and non-profit researchers from around the country to detail which diseases they’re trying to generate treatments for, so AtomNet can take a shot. The academic labs will receive 72 different drugs that the neural network has found to have the highest probability of interacting with the disease, based on the molecular data it’s seen.
Atomwise’s system only generates potential drugs—the compounds created by the neural network aren’t guaranteed to be safe, and need to go through the same drug trials and safety checks as anything else on the market. The company believes that the speed at which it can generate trial-ready drugs based on previous safe molecular interactions is what sets it apart.
Atomwise touts two projects that show the potential of AtomNet, drugs for multiple sclerosis and Ebola. The MS drug has been licensed to an undisclosed UK pharmacology firm, according to Atomwise, and the Ebola drug is being prepared for submission to a peer-reviewed publication.
Alexander Levy, the company’s COO and cofounder, said that AtomNet learns the interactions between molecules much like artificial intelligence learns to recognize images. Image recognition finds reduces patterns in images’ pixels to simpler representations, teaching itself the bounds of an idea like a horse or a desk lamp through seeing hundreds or thousands of examples.
“It turns out that the same thing that works in images, also works in chemistry,” Levy says. “You can take an interaction between a drug and huge biological system and you can decompose that to smaller and smaller interactive groups. If you study enough historical examples of molecules … and we’ve studied tens of millions of those, you can then make predictions that are extremely accurate yet also extremely fast.”
One of Atomwise’s advantages, Levy says, is that the network works with 3D models. To generate the drugs, the model starts with a 3D model of a molecule—for example a protein that gives a cancer cell a growth advantage. The neural network then generates a series of synthetic compounds (simulated drugs), and predicts how likely it would be for the two molecules to interact. If a drug is likely to interact with the specific molecule, it can be synthesized and tested.
Levy likens the idea to the automated systems used to model airplane aerodynamics or computer chip design, where millions of scenarios are mapped out within software that accurately represents how the physical world works.
“Imagine if you knew what a biological mechanism looked like, atom by atom. Could you reason your way to a compound that did the thing that you wanted?” Levy says.