Algorithmic Culture

Some excerpts from this article point out the effect machine algorithms have on shaping our information, our entertainment, and our culture.

The Creepy and Creeping Power of Social Media

By Ned Ryun| June 8th, 2018

While algorithms are necessary to serve up the content people want, social media companies failing to be transparent on this front are dangerous…Algorithm tweaking isn’t neutral and it has a massive “follow on” effect in the digital industry and political world, changing the kind of content that people see everyday. So if the algorithm starts filtering [say, content] it puts a thumb on the scale, favoring one side over the other. With a small handful of controllers over the algorithms, it’s appropriate to ask who controls the controllers?

….

We should acknowledge that rule by algorithm can be just as stringent as any rule by a dictator, perhaps even more so as it is vague, faceless, and hard to define. These algorithms decide what you see and don’t see in your timeline, subtly determining for you what is “worthy” of your attention. Facebook treats this algorithm like a black box, we’re never allowed to look inside and see what’s going on, we’ll only ever see the results on our news feeds. A world ruled by algorithms—just like the one it replaced controlled by network executives—closes off views, closes off debates, and further Balkanizes people. So, in fact, how can these social media and tech giants save democracy when in fact they’re becoming less democratic?

This is the same dynamic that is filtering and feeding our artistic content through the world-wide web. We can only consume the content that we can find and this is how it’s being found.

Put The Damn Phone Down and Do Something

This is a good interview with the Ben Silbermann, founder of Pinterest, published on Medium.

Some excerpts:

We’re social creatures. We need to connect with other people.

Pinterest is actually…it’s really about you. It’s about your tastes, your aspirations, your plans. There are other people there. Our recommendations are all curated by other users. The objective is not to do that [seek Likes]. That’s why it’s different than social networks.

sure, it’s fun to look at millions of ideas, but eventually, the real satisfaction and joy comes from giving it a shot. It might turn out great. It might turn out poorly. All of that is fine. We want to be the company that motivates you to put your phone down and to go try those things.

So, Pinterest is doing the right things to encourage engagement within the community. The next step of Web 3.0 is to distribute the network value they create back to the community of users. tuka will do that.

Surviving the Digital Economy

This will be a crucial issue for a free society going forward…giving away your data is like giving away your labor.

Want Our Personal Data? Pay for It

The posting, tagging and uploading that we do online may be fun, but it’s labor too, and we should be compensated for it

By Eric A. Posner and Glen Weyl

WSJ, April 20, 2018 11:19 a.m. ET

Congress has stepped up talk of new privacy regulations in the wake of the scandal involving Cambridge Analytica, which improperly gained access to the data of as many as 87 million Facebook users. Even Facebook chief executive Mark Zuckerberg testified that he thought new federal rules were “inevitable.” But to understand what regulation is appropriate, we need to understand the source of the problem: the absence of a real market in data, with true property rights for data creators. Once that market is in place, implementing privacy protections will be easy.

We often think of ourselves as consumers of Facebook, Google, Instagram and other internet services. In reality, we are also their suppliers—or more accurately, their workers. When we post and label photos on Facebook or Instagram, use Google maps while driving, chat in multiple languages on Skype or upload videos to YouTube, we are generating data about human behavior that the companies then feed into machine-learning programs.

These programs use our personal data to learn patterns that allow them to imitate human behavior and understanding. With that information, computers can recognize images, translate languages, help viewers choose among shows and offer the speediest route to the mall. Companies such as Facebook, Google, and Microsoft (where one of us works) sell these tools to other companies. They also use our data to match advertisers with consumers.

Defenders of the current system often say that we don’t give away our personal data for free. Rather, we’re paid in the form of the services that we receive. But this exchange is bad for users, bad for society and probably not ideal even for the tech companies. In a real market, consumers would have far more power over the exchange: Here’s my data. What are you willing to pay for it?

An internet user today probably would earn only a few hundred dollars a year if companies paid for data. But that amount could grow substantially in the coming years. If the economic reach of AI systems continues to expand—into drafting legal contracts, diagnosing diseases, performing surgery, making investments, driving trucks, managing businesses—they will need vast amounts of data to function.

And if these systems displace human jobs, people will have plenty of time to supply that data. Tech executives fearful that AI will cause mass unemployment have advocated a universal basic income funded by increased taxes. But the pressure for such policies would abate if users were simply compensated for their data.

The data currently compiled by Facebook and other companies is of pretty low quality. That’s why Facebook has an additional army of paid workers who are given dedicated tasks, such as labeling photos, to fill in the gaps left by users. If Facebook paid users for their work, it could offer pay tied to the value of the user’s contribution—offering more, for example, for useful translations of the latest Chinese slang into English than for yet another video labeled “cat.”

So why doesn’t Facebook already offer wages to users? For one, obviously, it would cost a lot to pay users for the data that the company currently gets for free. And then Google and others might start paying as well. Competition for users would improve the quality of data but eat away at the tech companies’ bottom line.

It’s also true that users simply aren’t thinking this way. But that can change. The basic idea is straightforward enough: When we supply our personal data to Facebook, Google or other companies, it is a form of labor, and we should be compensated for it. It may be enjoyable work, but it’s work just the same.

If companies reject this model of “data as labor,” market pressure could be used to persuade them. Rather than sign up directly with, say, Facebook, people would sign up with a data agent. (Such services, sometimes referred to as personal data exchanges or vaults, are already in development, with more than a dozen startups vying to fill this role.) The data agent would then offer Facebook access to its members and negotiate wages and terms of use on their behalf. Users would get to Facebook through the agent’s platform. If at any time Facebook refused reasonable wages, the data agent could coordinate a strike or a boycott. Unlike individual users, the data agent could employ lawyers to review terms and conditions and ensure that those terms are being upheld.

With multiple data agents competing for users’ business, no one could become an abusive monopolist. The agent’s sole purpose would be managing workers’ data in their interests—and if there were a problem, users could move their data to another service without having to give up on their social network.

Companies such as Apple and Amazon also could get into the act. Currently, their business models are very different from those of Facebook and Google. For the most part, their focus is on selling products and services, rather than offering them without charge. If Facebook and Google refuse to pay users for their data, these other companies are big and sophisticated enough to pay for data instead.

Would the “data as labor” model put the tech giants out of business? Hardly. Their vast profits already reflect their monopoly power. Their margins would certainly be tighter under this new regime, but the wider economy would likely grow through greater productivity and a fairer distribution of income. The big companies would take a smaller share of a larger pie, but their business model would be far more sustainable, politically and socially. More important, they would have to focus on the value that their core services bring to consumers, rather than on exploiting their monopoly in user data.

As for Congress, it could help by making it simpler for individuals to have clear property rights in their own data, rights that can’t be permanently signed away by accepting a company’s confusing terms and conditions. The European Union has already taken steps in this direction, and its new regulations—which require data to be easily portable—are a leading stimulus for the rise of data agent startups. Government can also help by updating labor law to be more consistent with modern data work while protecting data workers from exploitation.

Most of us already take great satisfaction in using social media to connect with our friends and family. Imagine how much happier and prouder we would be if we received fair pay for the valuable work we perform in doing that?

Prof. Posner teaches at the University of Chicago Law School. Dr. Weyl teaches at Yale University and is a principal researcher at Microsoft (whose views he in no way represents here). Their new book is “Radical Markets: Uprooting Capitalism and Democracy for a Just Society,” which will be published on May 8 by Princeton University Press.

https://www.wsj.com/articles/want-our-personal-data-pay-for-it-1524237577 (Paywall)

The Real Scandal?

likenolike

The Real Scandal Isn’t What Cambridge Analytica Did

It’s what Facebook made possible.

A couple of excerpts:

Sinister as it sounds, “psychographic” targeting—advertising to people based on information about their attitudes, interests, and personality traits—is an imprecise science at best and “snake oil” at worst.

….

If you think of that data, and the ads, as a relatively small price to pay for the privilege of seamless connection to everyone you know and care about, then Facebook looks like the wildly successful, path-breaking company that made it all possible. But if you start to think of the bargain as Faustian—with hidden long-term costs that overshadow the obvious benefits—then that would make Facebook the devil.

What this scandal did, then, was make the grand bargain of the social web look a little more Faustian than it did before.

From that perspective, the real scandal is that this wasn’t a data breach or some egregious isolated error on Facebook’s part. What Cambridge Analytica did was, in many ways, what Facebook was optimized for—collating personal information about vast numbers of people in handy packets that could then be used to try to sell them something.

Yes, the real scandal is that most of us are giving away real value that we need to survive in the digital economy.

Who’s Watching You?

Personal data is the new goldmine. Here’s who’s mining your gold:

What You Need to Know About Tech

Good article.

12 Things Everyone Should Understand About Tech

A couple of good excerpts:

9. Most big tech companies make money in just one of three ways.

It’s important to understand how tech companies make money if you want to understand why tech works the way that it does.

  • Advertising: Google and Facebook make nearly all of their money from selling information about you to advertisers. Almost every product they create is designed to extract as much information from you as possible, so that it can be used to create a more detailed profile of your behaviors and preferences, and the search results and social feeds made by advertising companies are strongly incentivized to push you toward sites or apps that show you more ads from these platforms. It’s a business model built around surveillance, which is particularly striking since it’s the one that most consumer internet businesses rely upon.
  • Big Business: Some of the larger (generally more boring) tech companies like Microsoft and Oracle and Salesforce exist to get money from other big companies that need business software but will pay a premium if it’s easy to manage and easy to lock down the ways that employees use it. Very little of this technology is a delight to use, especially because the customers for it are obsessed with controlling and monitoring their workers, but these are some of the most profitable companies in tech.
  • Individuals: Companies like Apple and Amazon want you to pay them directly for their products, or for the products that others sell in their store. (Although Amazon’s Web Services exist to serve that Big Business market, above.) This is one of the most straightforward business models—you know exactly what you’re getting when you buy an iPhone or a Kindle, or when you subscribe to Spotify, and because it doesn’t rely on advertising or cede purchasing control to your employer, companies with this model tend to be the ones where individual people have the most power.

That’s it. Pretty much every company in tech is trying to do one of those three things, and you can understand why they make their choices by seeing how it connects to these three business models.


10. The economic model of big companies skews all of tech.

Today’s biggest tech companies follow a simple formula:

  1. Make an interesting or useful product that transforms a big market
  2. Get lots of money from venture capital investors
  3. Try to quickly grow a huge audience of users even if that means losing a lot of money for a while
  4. Figure out how to turn that huge audience into a business worth enough to give investors an enormous return
  5. Start ferociously fighting (or buying off) other competitive companies in the market

This model looks very different than how we think of traditional growth companies, which start off as small businesses and primarily grow through attracting customers who directly pay for goods or services. Companies that follow this new model can grow much larger, much more quickly, than older companies that had to rely on revenue growth from paying customers. But these new companies also have much lower accountability to the markets they’re entering because they’re serving their investors’ short-term interests ahead of their users’ or community’s long-term interests.

The pervasiveness of this kind of business plan can make competition almost impossible for companies without venture capital investment. Regular companies that grow based on earning money from customers can’t afford to lose that much money for that long a time. It’s not a level playing field, which often means that companies are stuck being either little indie efforts or giant monstrous behemoths, with very little in between. The end result looks a lot like the movie industry, where there are tiny indie arthouse films and big superhero blockbusters, and not very much else. [This is amplifying the winner-take-all dynamics of technology.]

And the biggest cost for these big new tech companies? Hiring coders. They pump the vast majority of their investment money into hiring and retaining the programmers who’ll build their new tech platforms. Precious little of these enormous piles of money is put into things that will serve a community or build equity for anyone other than the founders or investors in the company. There is no aspiration that making a hugely valuable company should also imply creating lots of jobs for lots of different kinds of people. [Note: I would add to this last point that the aspiration should be to distribute value more widely across the community, not necessarily create more jobs.]

What are the DApps of the Future?

But how?

DApps, or Distributed Applications, are the force multipliers for blockchain technologies, just like email, Amazon, eBay, Google, and social networks are the applications that have propelled the Internet. The race is on for the development of these Dapps to transform industries and the future of the Internet itself.

In Search of Blockchain’s Killer-Apps

By Irving Wladawsky-Berger, WSJ, Mar 9, 2018

Blockchain has been in the news lately, but beyond knowing that it has something to do with payments and digital currencies, most people don’t know what blockchain is or why they should care. A major part of the reason is that we still don’t have the kind of easy-to-explain blockchain killer-apps that propelled the internet forward.

Blockchain has yet to cross the chasm from technology enthusiasts and visionaries to the wider marketplace that’s more interested in business value and applications. There’s considerable research on blockchain technologies, platforms and applications as well as market experimentation in a number of industries, but blockchain today is roughly where the internet was in the mid-late 1980s: full of promise but still confined to a niche audience.

In addition, outside of digital currencies, blockchain applications are primarily aimed at institutions. And, given that blockchain is all about the creation, exchange and management of valuable assets, its applications are significantly more complex to understand and explain than internet applications.

The management of information is quite different from the management of transactions. The latter, especially for transactions dealing with valuable or sensitive assets, requires deep contractual negotiations among companies and jurisdictional negotiations among governments. Moreover, since blockchain is inherently multi-institutional in nature, its applications involve close collaboration among companies, governments and other entities.

In my opinion, there will likely be two major kinds of blockchain killer-apps: those primarily aimed at reducing the friction and overheads in complex transaction involving multiple institutions; and those primarily aimed at strengthening the security and privacy of the internet through identity management and data sharing. Let me discuss each in turn.

Complex transactions among institutions. “Contracts, transactions, and the records of them are among the defining structures in our economic, legal, and political systems,” wrote Harvard professors Marco Iansiti and Karim Lakhani in a 2017 HBR article.

With blockchain, “every agreement, every process, every task, and every payment would have a digital record and signature that could be identified, validated, stored, and shared… Individuals, organizations, machines, and algorithms would freely transact and interact with one another with little friction.”

Blockchain holds the promise to transform the finance industry and other aspects of the digital economy by bringing one of the most important and oldest concepts, the ledger, to the internet age. Ledgers constitute a permanent record of all the economic transactions an institution handles, whether it’s a bank managing deposits, loans and payments; a brokerage house keeping track of stocks and bonds; or a government office recording the ownership and sale of land and houses.

Over the years, institutions have automated their original paper-based ledgers with sophisticated IT applications and data bases. But while most ledgers are now digital, their underlying structure has not changed. Each institution continues to own and manage its own ledger, synchronizing its records with those of other institutions as appropriate, – a cumbersome process that often takes days. While these legacy systems operate with a high degree of robustness, they’re rather inflexible and inefficient.

In August of 2016, the WEF published a very good report on how blockchain can help reshape the financial services industry. The report concluded that blockchain technologies have great potential to drive simplicity and efficiency through the establishment of new financial services infrastructure, processes and business models.

However, transforming the highly complex global financial ecosystem will take considerable investment and time. It requires the close collaboration of its various stakeholders, including existing financial institutions, fintech startups, merchants of all sizes, government regulators in just about every country, and huge numbers of individuals around the world. Getting them to work together and pull in the same direction is a major undertaking, given their diverging, competing interests. Overcoming these challenges will likely delay large-scale, multi-party blockchain implementations.

Supply chain applications will likely be among the earliest blockchain killer-apps, increasing the speed, security and accuracy of financial and commercial settlements; tracking the supply chain lifecycle of any component or product; and securely protecting all the transactions and data moving through the supply chain. The infrastructures and processes of supply chains are significantly less complex than those in financial services, healthcare, and other industries and there are already a number of experimental applications under way.

A recent WSJ CIO Journal article noted that blockchain seems poised to change how supply chains work. The article cites examples of projects with Walmart and British Airwayswhere blockchain is used to maintain the integrity of the data being shared across the various institutions participating in their respective ecosystems. Earlier this year IBM and Maersk announced a joint venture to streamline operations for the entire global shipping ecosystem. Their joint venture aims to apply blockchain technologies to the current stack of paperwork needed to process and track the shipping of goods. Maersk estimates that the costs to process and administer the required documentation can be as high as 20 percent the actual physical transportation costs.

Identity management and data sharing. The other major kind of blockchain killer-apps will likely deal with identity management and data security.

As we move from a world of physical interactions and paper documents, to a world primarily governed by digital data and transactions, our existing methods for protecting identities and data are proving inadequate. Internet threats have been growing. Large-scale fraud, data breaches, and identity thefts are becoming more common. Companies are finding that cyberattacks are costly to prevent and recover from. The transition to a digital economy requires radically different identity systems.

A major reason for the internet’s ability to keep growing and adapting to widely different applications is that it’s stuck to its basic data-transport mission.  Consequently, there’s no one overall owner responsible for security, let alone identity management, over the internet. These important responsibilities are divided among several actors, making them significantly harder to achieve.

Blockchain technologies should help us enhance the security of digital transactions and data, by developing the required common services for secure communication, storage and data access, along with open source software implementations of these standard services, supported by all major blockchain platforms, such as Hyperledger and Ethereum.

Identity is the key that determines the particular transactions in which individuals, institutions, and the exploding number of IoT devices, can rightfully participate, as well as the data they’re entitled to access. But, our existing methods for managing digital identities are far from adequate.

To reach a higher level of privacy and security we need to establish a trusted data ecosystem, which requires the interoperability and sharing of data across the various institutions involved. The more data sources a trusted ecosystem has access to, the higher the probability of detecting fraud and identity theft. However, it’s not only highly unsafe, but also totally infeasible to gather all the needed attributes in a central data warehouse. Few institutions will let their critical data out of their premises.

MIT Connection Science, a research initiative led by MIT professor Sandy Pentland, has been developing a new identity framework that would enable the safe sharing of data across institutions. Instead of copying or moving the data across, the agreed upon queries are sent to the institution owning the data, executed behind the firewalls of the data owners, and only the encrypted results are shared. MIT Connection Science is implementing such an identity framework in its OPAL initiative, which makes extensive use of cryptographic and blockchain technologies. A number of pilots are underway around the world.

Irving Wladawsky-Berger worked at IBM for 37 years and has been a strategic advisor to Citigroup and to HBO. He is affiliated with MIT, NYU and Imperial College, and is a regular contributor to CIO Journal.

Who Are We and What Matters?

Google? Facebook? These two firms alone control roughly 2/3s of digital media advertising revenues.

That’s power. That’s knowledge.

But knowledge of what?

Mostly of how to program computers and deploy algorithms to sort through, organize, cluster, rank, and order vast quantities of data. In the case of Facebook, Zuckerberg obviously also understood something simple but important about how human beings might enjoy interacting online. That’s not nothing. Actually, it’s a lot. An enormous amount. But it’s not everything — or anything remotely close to what Silicon Valley’s greatest innovators think it is.

When it comes to human beings — what motivates them, how they interact socially, to what end they organize politically — figures like Page and Zuckerberg know very little. Almost nothing, in fact. And that ignorance has enormous consequences for us all.