The search for Artificial General Intelligence

We already have some pretty fine examples of AI, but all of them are limited to performing a specific task. However, the ultimate search is for Artificial General Intelligence (AGI), which is effectively a complete ‘human-like’ AI system. This would be a system that basically can think like a human. But, as Dan Robitski wrote in futurism.com, “No amount of optimizing systems to get better at a particular task would ever lead to AGI.” In other words, he doesn’t think we’re going to just stumble across it. The reason being; companies working on AI systems have a narrow focus.

What would be the benefits of AGI?

If we had a system that was able to use abstract reasoning and everything that goes with that, including creativity, we would be able to make a major leap in solving the problems associated with aspects of space exploration, economics, healthcare and much more. There are lots of aspects of our lives that AGI could positively impact on.

AGI would also be a massively interesting investment and a very valuable one. However, Robitski believes that investment money would need to come from governments rather than private investors and venture capital, simply because the way private funds structure their businesses would mean that a development platform would never be able to raise enough money.

Why private investors are avoiding AGI

Marian Gazdik, Managing Director of Startup Grind Europe said: “Investors only fund something when they see the end of the tunnel, and in AI it’s very far.” Hence the need for government funds. Tak Lo, a partner at Zeroth.ai, an Asian accelerator that invests in technology startups, who was speaking alongside Gazdik at The Joint Multi-Conference on Human-Level Artificial Intelligence in Prague last week also commented: “I very much like General AI as an intellectual, but as an investor not as much.”

Venture capitalists like Lo prefer to “invest in companies with great business models that use AI to solve a big problem, or companies that got their hands on a large, valuable dataset for training algorithms.”

The problem with AGI is that a workable solution is so far away in the future that private investors just can’t see a time when they’ll see a return. They think in terms of five to ten years, and we may have to wait longer than that before AGI becomes a reality. On the other hand a government doesn’t have to be tied to the same time schedule: they can put money into a project that serves the greater good without being too concerned about when the end product is delivered. But, first they have to decide that AGI is a project they want to get behind and so far no major government has made that commitment. Until they do, AGI will remain a dream.

Bill Gates: economics fit for a tech future

Basic economics always taught us that ‘supply and demand’ was a central feature of understanding a market. However, Bill Gates wrote in his blogrecently that “supply and demand is over” and he argued that it simply doesn’t apply to today’s economy. He also stated that politicians aren’t paying enough attention to this economic shift.

Why does he make this claim?

His reasoning is based on the fact that companies are no longer only making money by selling tangible products. Companies that supply software being one example. To develop new software, Gates points out, all of the cost is upfront, whereas a traditional manufacturer has to pay for parts and labour. When Microsoft launches a new version of a software programme, it can be copied, sold, and downloaded indefinitely for the relatively minimal costs of distribution and server space.

Gates claims that more large companies are operating without tangible products and says that digital products, which are a so-called “intangible investment,” carry new risks for businesses and investors and that this is not being accounted for in economic thinking, which still relies too much on an old model.

Capitalism without Capital

In his book “Capitalism without Capital” Gates presents the idea that developing software is a “sunk cost” because developers can’t recoup their losses the way other companies might. If you manufacture tangible products and go bust, you can sell off machinery, but a tech company doesn’t have any such assets to sell.

Gates also points out, the Gross Domestic Product (GDP), the sum of all goods and services sold in a country that is often used as a benchmark for an economy’s well-being doesn’t factor in the investment in intangible elements needed to make a product marketable, such as research and development or market research. He also suggests that didn’t matter two decades ago, but now it does because tech companies make up a bigger slice of a country’s GDP these days. And governments haven’t caught up with this fact.

Gates doesn’t offer a new economic model, but as he says: “The idea today that anyone would need to be pitched on why software is a legitimate investment seems unimaginable, but a lot has changed since the 1980s. It’s time the way we think about the economy does, too.”

 

AI enhanced humans

Could you use an extra hand?

Somebody may have said to you at sometime: “I’ve only got two hands,” indicating that whatever you asked them to do just isn’t possible. But in the field of prosthetics and what are called “enhanced humans” it may well be possible now to have additional limbs to carry out tasks.

We have devices that wearers can control with their thoughts — these help people with prosthetic limbs to feel more whole, but now researchers are setting out to see if such devices could make humans more than whole.

Advanced Telecommunications Research Institute in Japan wanted to know if giving someone a supernumerary robotic limb (SRL), a mind-controlled robotic limb that worked alongside the person’s two biological ones, could give that person multitasking abilities beyond those of the average human.

The study

The researchers asked 15 volunteers to sit in a chair with an SRL positioned as if it were a third arm coming from their own body. Each volunteer wore a cap that that tracked the brain’s electrical activity and the cap transmitted this activity to a computer that then turned it into movement in the SRL. All the volunteers had to do to get the SRL to work was think about what they wanted it to do.

The next stage was to ask the volunteers to complete two tasks. The first one — balancing a ball on a board –they did using their natural arms and hands. The second one — picking up and putting down a bottle –they did using the SRL. They then asked the volunteers to do both the tasks separately and simultaneously.

The results

The results of 20 trials showed that the volunteers were able to complete both tasks successfully using the three limbs about 75 percent of the time. What this means is that they were able to complete two tasks simultaneously that they couldn’t have done if they had been limited to using their natural limbs.

The researchers also think that by operating this brain-machine interface, we have an idea that we may be able to train the brain itself and their future research will try to establish if we might be able to enhance our minds by temporarily enhancing our bodies.

Why you need a decentralised identity

We have all been warned about identity theft. Even the big banks like Barclays are running TV ad campaigns showing customers how a phone call that seems to legitimately come from your bank can be used to steal your online banking pin number. There is plenty of information out there about how to keep your details safe, but no matter what precautions we take, there are always bad actors out there (this is now a ‘polite’ way of referring to people who are nothing more than criminals) who are relentless in their search for new ways to access our private information.

Tomislav Markovski, writing on Medium tells a story about how he nearly became the victim of bank fraud when he rented a property. After providing every possible kind of document to the real estate agent, including bank statements and investment portfolio details, he received a call from his bank a few days after he had moved in saying that someone wanted to cash a large check drawn on his account. Markovski knew he didn’t have that much money in is account, but the bank then told him that “he” had made a transfer from his savings account by phone. Of course, he’d made no such call, and thanks to his bank calling when they did, the theft was stopped. But, as he says, it was a “masterfully crafted plan that involved just four key steps”

1. Call the bank pretending to be Markovski

2. Change his phone number (to confirm large withdrawal)

3. Transfer all his savings into his current account

4. Have a fake cheque made and present it to the bank for withdrawal

They were able to do this because they had access to all the necessary information on him, including his social security number. They couldn’t catch the scammer, but it made Markovski think about why so much information was required to rent an apartment and why are we still relying on physical documents.

Blockchain has a solution — decentralised identity

Blockchain technology is opening up a range of possibilities to prevent this kind of crime and decentralised identity could be the way forward. As Markovski says, decentralised identity is “publicly discoverable identity information.” It uses blockchain technology to provide tamper-evident information about an entity or a subject and “allows a model of truth to be established between parties that rely on communication and exchange of data.”

There are already a few platforms working on this, including Civic, uPort and Sovrin. As Markovski says: “Decentralized identity platforms will change the current broken identity system that relies on numerous online services requiring us to remember passwords for each of them. They can help us protect our personal information and allow us to control how this data is shared.”

Until these platforms gain mass adoption — be careful out there!