There's no small amount of controversy between those who feel that artificial intelligence is an essential part of the future and those who say that artificial intelligence will kill us all. Let's consider those divergent points of view.
British science fiction writer Arthur C. Clarke formulated three adages that are known as Clarke's three laws starting in 1963 with "When a distinguished but elderly scientist states that something is possible, he is almost certainly right. When he states that something is impossible, he is very probably wrong." We're certainly seeing that with the advances being made in artificial intelligence.
Later he added the second law, "The only way of discovering the limits of the possible is to venture a little way past them into the impossible." And around 1972, he added the third, "Any sufficiently advanced technology is indistinguishable from magic."
Some scientists and technologists are offended by that third law, apparently because they have misread it and believe that Clarke was saying that technology is magic. Parsing the sentence shows that this is not the case at all. Had Clarke wanted to say that technology is magic, his third law would have been "any sufficiently advanced technology is magic" instead of "any sufficiently advanced technology is INDISTINGUISHABLE FROM magic."
In more than a few ways, Adobe's Sensei technology has reached the "indistinguishable from magic" plateau. Many of the advanced features added to the company's various products in the past couple of years are dependent on Sensei. Writing on the Adobe Blog, Scott Prevost, a vice president of engineering at Adobe, says "With artificial intelligence (AI) and machine learning, computers today learn as they are exposed to more data — without the need for explicit programming. They’re really good at finding and understanding patterns in data and inferring rules about it."
But that's really just the beginning. Today technologies such as Sensei use "content understanding" to help digital artists locate appropriate images. This is seen primarily in Adobe's stock image offerings where key words can be combined with other aspects such as depth of field, color intensity, and image content to shorten the time needed to find the right image.
Then the "content aware" functions come into play. Someone who is creating an advertisement may have identified the ideal image, except that it has too much space or too little space for text. The various content aware functions, and particularly the most recent "select subject" option can save a lot of time as the graphic designer modifies the image to fit the need.
A third function Prevost mentions in his blog post is the ability for machine learning to assist the designer in evaluating and predicting the effectiveness of the work. I'm familiar with the first two uses, but this third one stopped me. We're definitely encroaching on the area covered by Clarke's Third Law now.
Here's what Provost wrote in explaining how AI and machine learning can make it possible for brands and businesses to personalize messages and measure how consumers respond to them: "[M]achine learning helps make predictions and recommendations about which combinations of images, video, and text will drive the most on-line traffic or generate the most conversions for particular audience segments. This could help a marketer craft a different creative for users on mobile compared to desktop users, easily switching out images and messaging to align with what the audience expects to engage with on that channel."
Artificial intelligence is what makes it possible to find a store that sells what you're looking for. AI is what your bank uses to send a text when an unusual charge appears on your credit card. It's what helps you remove an object or person from the background of a photo you took.
If you'd like to learn more about how it's possible to create a machine that thinks, take a look at an introduction to artificial intelligence on the Adobe blog.
The disruption and transformation that will result from AI are remarkable. In fact, that is likely part of what attracted Elon Musk's attention. The man responsible for Space-X, the Tesla electric car, and some real breakthroughs in solar power detests artificial intelligence.
Vanity Fair reports an incident in which Musk and AI developer Demis Hassabis were discussing the most important projects in the world. Musk reportedly said that the ultimate goal of Space-X is interplanetary colonization, in part to have an escape plan when AI takes over. Hassabis said that AI would simply follow to Mars.
Musk has been promoting the anti-AI message since 2014 and he maintains that it is the greatest existential threat to human life. Musk says that the leading AI companies have tried to ensure safety because they recognize the danger. AI developers "believe that they can shape and control the digital super-intelligences and prevent bad ones from escaping into the internet," Musk wrote. "That remains to be seen."
We can hope Musk is wrong about that, but if the past is any indicator of the future, good technology can be put to bad use -- and has been time after time.
Computers have been able to pass Alan Turing's 1951 test for quite some time and modern AI technology really got underway the 1980s when if-then logic rules allowed what were at the time called "expert systems" to perform basic reasoning. These step-by-step systems were cumbersome to build and were likely to fail when the system encountered a condition that the programmers hadn't accounted for.
Now computer scientists are developing neural networks. A designer who needs to remove something from a photograph using Photoshop has several tools to choose from. If the needed change is something simple -- an airplane in a blue sky, for example -- the clone stamp tool is usually sufficient. On the other hand, a more complex change such removing a tree from an image that contains several objects in the background and a sky with clouds might be handled better by content-aware fill that makes extensive use of AI.
One thing is clear: AI is going to be a part of our future and we probably shouldn't be too surprised when we encounter unintended consequences. This is probably not a good time to think about "The Sorcerer's Apprentice" from Disney's Fantasia.
It's not uncommon to hear people say that they would drop Facebook immediately if a viable alternative existed, but one doesn't. Perhaps the latest news about Facebook and Cambridge Analytica will at least convince people to use Facebook more safely. The Cambridge Analytica chapter is only the latest chapter in a long story.
The trouble with opting out of Facebook entirely is that there's currently no comparable replacement to share information with family and friends (and political rivals, of course). Some have suggested an optional annual fee that would allow users to opt out of all features that Facebook uses to make money. You'd have the choice of paying some fee for the service without ads and other unwanted "features" or continuing to use the service for free and accept the fact that you are Facebook's product and advertisers are their real customers.
There are three actions you can take right now to help:
Matt Kruse, the developer of Social Fixer, says we should ignore Facebook "quizzes" and "personality tests" and stop using "those cute Most Used Words apps." Those apps harvest more of your data than most of us realize and often they continue to read all your posts long after you've forgotten about the app. In some cases, they can see data from your friends, too.
The Electronic Frontier Foundation has an post that explains how to opt out of platform sharing. It's an easy process. Just log in to Facebook and type this in the browser's address bar: https://www.facebook.com/settings?tab=applications. Then edit "Apps, Websites, and Plugins" to disable the feature.
If you don't want to entirely disable the platform apps, open Apps Others Use and then deselect the types of information that you want to hide from the apps.
I use and recommend AdBlock Plus, a free plug-in for most browsers. You'll find it in the browser's add-ons or extensions section.
Social Fixer for Facebook is my preference, but Facebook Purity (also known as "Fluff Buster") is also a good choice. Both of these are available from the browser's add-ons or extensions section. Facebook Purity was developed by Steve Fernandez, a UK-based programmer, as donation-ware. It is available for Firefox, Chrome, Microsoft Edge, Safari, Opera, and Maxthon. Social Fixer by Matt Kruse is also donation-ware. and runs on Firefox, Chrome, Safari, and Opera.
This week, investors reduced the value of Facebook by about $50 billion and politicians in the US and Europe are calling for investigations, including a demand that CEO Mark Zuckerberg testify to investigators.
There's increased attention from the Federal Trade Commission because Facebook accepted a settlement in 2011 in which it promised that third-party apps would not be allowed to access users' data. In fact, little changed and the apps have continued to be able to access a great deal of data about users. The FTC could fine Facebook up to $40,000 per day, which probably seems like a lot of money to you, but Facebook has daily revenues of more than $100 million so the fine would be paid out of petty cash.
Businesses and government agencies in the US will spend $553 billion this year on research and development. Innovation is one of the indicators that Amazon is considering in its search for a second headquarters location. Research that ranks the states based on innovation has some surprises.
But first, consider Amazon's list: Columbus, where I live, surprised a lot of people by showing up in Amazon's list of 20 contenders. It's an unlikely prospect. The other 19 are Atlanta, Austin, Boston, Chicago, Dallas, Denver, Indianapolis, Los Angeles, Miami, Montgomery County (Maryland), Nashville, Newark, New York City, Northern Virginia, Philadelphia, Pittsburgh, Raleigh, Toronto, and Washington.
If innovation is one of Amazon's primary considerations, it's not good news that Ohio ranks 33rd on WalletHub's analysis that takes into account 22 key indicators of innovation friendliness such as science, technology, engineering, and math professionals; research and development spending per capita; and tech-company density. According to WalletHub, Massachusetts is the most innovative state and Mississippi is the least innovative.
Ohio is one step above Alaska, but below states such as Montana, Indiana, Wyoming, and Michigan. Taking a wider view, the United States is the 4th most innovative nation on the planet, falling behind Switzerland, Sweden, and Netherlands. Togo, Guinea, and Yemen are at the bottom of the world list according to the Global Innovation Index.
The WalletHub research compared the 50 states and the District of Columbia across two key dimensions, human capital and innovation environment. The full report and an explanation of the methodology are on the WalletHub website.
The Slow ring computer on my desk received an update to Windows version 17115 this week and it previews what users will see in a few weeks when Microsoft issues the Spring update for Windows 10. The coming update will include changes to the setup experience for privacy settings in a way that Microsoft hopes will allow users to make better decisions about computer privacy. There will also be new settings for Inking & Typing and Find My Device.
These settings may appear on a single screen or on separate screens in the preview version. Both of those options have been pushed out, each to a different set of insider users.
The settings will allow users to have more granular control over what information is sent to Microsoft.
For Inking and Typing, users can allow the system to send data to Microsoft so that it can be used to improve language recognition. Developers use the information to find ways to make the process better and to improve predictive text suggestions.
The Find My Device option is useful for all computers, but particularly for mobile devices. When activated, the service can report the location of the computer to the owner. This can be helpful in attempting to recover a stolen computer. It works only if the user has created a Microsoft account, not just a local account on the computer.
So the Spring update will be a low-key event, but don't expect any groundbreaking changes this time around.
