Tag: data

DALL-E! AI-Generated Art- Who owns the AI art you are making?

Tony DePrato2 Comments

I attended an AI talk last week. The presenter was an AWS Engineer (Amazon Webservices). They sit on a few committees and advisory counsels for the current administration in Washington, D.C.

The talk was a high-level view of how the technology works. Towards the end, a person in the audience asked, “Who owns the copyright on images produced by AI art?”

The answer: Not You.

This ownership idea connects to other conversations I have had about the terms of service with ChatGPT and my initial chat with the AI about how it handles data.

This technology needs new laws and structures around it, and those will come with lawsuits and public debate. As of now, it is not ready for children to use in a K-12 environment. It is not ready for confidential information and data handling (unless a paid option includes the ability to label data).

In terms of the art, and I assume audio and other media it creates, it is not ready to be used in a business or academic environment where ownership is required.

Image using the DALL-E! system, and embedding the art on your organization’s website. You do this routinely for 2-3 years. Then, one day, you are notified that the copyright owner is seeking justification for using the content in a commercial setting and would like to license the art.

Maybe you are wondering, how can they do this? How can they tag the art and track it years later? There are numerous ways to do this, but if they have an AI that can create art, they have an AI that can track art using techniques that would be invisible to the end user. If I were working at OpenAI, I would develop a steganographic method to find content in the future. This technique would place content hidden within the content on the image. It could be achieved using ratios and distance, allowing even the smallest thumbnail to be discovered.

That’s just one idea from someone who thought about it for a day. Image a team of people who built an AI.

The general rule for operating a business: don’t break the law

If you work at a school or a bank, reading the terms of service concerning minors, confidential data, and copyright is not an option. If you are determined to push forward, have your organization’s legal team review what you are planning, even webinars about advocating use.

Here is my chat with the OpenAI Chatbot regarding how copyright is currently handled (or not handled):

Are familiar with an AI that can generate Art?
Yes, I am familiar with AI that can generate art. There are a number of different approaches to this, but some of the most popular ones include:

Generative Adversarial Networks (GANs): These are a type of AI that are trained by having two neural networks compete against each other. One network, called the generator, creates images, while the other network, called the discriminator, tries to distinguish between the real and fake images. Over time, the generator gets better and better at creating images that the discriminator can’t distinguish from real ones.

Variational Autoencoders (VAEs): These are a type of AI that are trained to learn the underlying structure of a dataset, such as images of faces. Once the VAE has learned this structure, it can generate new, similar images by sampling from the learned distribution.

Style transfer: This is a type of AI that takes two images — one as the “content” and one as the “style” — and blends them together to create a new image that has the content of one and the style of the other.

There are many other approaches to AI-generated art, but these are some of the most popular.

anthony.deprato@gosaints.org
Are you familiar with Dall-E
Yes, I am familiar with DALL·E! DALL·E is an AI-powered image synthesizer created by OpenAI. It was introduced in January 2021 and quickly gained attention for its ability to generate a wide variety of images from textual descriptions, including many that are fantastical or imaginative in nature. For example, DALL·E can generate images of objects that don’t exist in the real world, such as a hotdog-shaped bicycle or a horse with a flashlight on its head. The images generated by DALL·E are unique and often quite striking, which has led to a lot of interest and excitement in the field of AI art and creativity.

anthony.deprato@gosaints.org
What company, person, or entity holds the copyright for DALL-E! generated art?
The ownership of the copyright for DALL-E-generated art is governed by the terms and conditions set forth by OpenAI, the company that created DALL-E. OpenAI holds the copyright to the DALL-E software, and it is likely that the images generated by the software are also considered to be works created by OpenAI for copyright purposes.

However, it is worth noting that the exact ownership of the copyright for AI-generated art is a complex issue that is still being debated in legal circles. There is no clear consensus on whether AI-generated works are considered to be the product of human authorship and therefore eligible for copyright protection, or whether they are considered to be works created by a machine and therefore not eligible for copyright protection.

In the case of DALL-E, OpenAI may choose to license the use of the images generated by the software to third parties, or it may choose to allow free use of the images with certain conditions. The specific terms and conditions for the use of DALL-E-generated art will depend on the policies and decisions of OpenAI.

anthony.deprato@gosaints.org
Does OpenAI use steganography to track DALL-E! art?
I’m not aware of OpenAI using steganography to track DALL-E-generated art. Steganography is a technique for hiding data within other data, such as embedding a hidden message within an image file. To my knowledge, OpenAI does not use steganography for tracking DALL-E-generated art.

It is possible that OpenAI may use other methods for tracking the use of DALL-E-generated art, such as watermarking the images or tracking usage through digital means, but I don’t have any specific information on the methods used by OpenAI. Ultimately, the method used for tracking the use of DALL-E-generated art will depend on OpenAI’s policies and decisions.

anthony.deprato@gosaints.org
Thank you
You’re welcome! If you have any other questions, feel free to ask.

Episode 173 – Colts for the Win!

Patrick CauleyLeave a comment

Tony and Patrick are back for another great episode. It is a quick one this week where we talk a little NFL, Google Daydream and swimming in your data. Check out the talking points below and as always be sure to subscribe to us on iTunes or your favorite podcasting app.

  1. Google Daydream is winding down
    1. https://www.androidauthority.com/google-daydream-1041548/
    2. https://killedbygoogle.com 
    3. https://www.oculus.com/medium/
  2. Swimming in the Data Lake by Tony DePrato
    1. https://itbabble.com/2019/10/12/swimming-in-the-data-lake/
  3. SideCar – Full Test…with Apple TV
    1. https://support.apple.com/en-us/HT210380
    2. Must use a pen
    3. Astropad  – https://astropad.com 

You can download this week’s episode HERE!

 

Swimming in the Data Lake

Tony DePrato1 Comment

hacker-2883632_960_720

By: Tony DePrato | Follow me on LinkedIn

Educational organizations are face with the constant influx of seemingly new technology. This creates pressure (even a demand) to review, to compare, to challenge, and often, to change.

The business around software is not understood nor recognized by most educational institutions. Schools trust. Schools hope. Industry, generally, tries to profit. The constant churning of technology companies through mergers and venture capital initiatives creates an unstable environment. This environment breeds havoc and forces decision makers to ask: Will your vendor be your vendor tomorrow?

The game is rigged. But. There is way to take the game back. There is way to design flexibility into the ecosystem and empower decision makers to relentlessly negotiate for the deals they need, and when they need them.

For years we have focused on controlling the box, the physical interface, and the platform. In most cases, these concepts are now irrelevant. Software as a Service (SAS) has evolved to allow a tablet to carry the power of a laptop, and a laptop the power of a GPU driven workstation. SAS is just the beginning, in fact, it is only a replacement for the platform. The heart of the system is the data, and the data controls the decisions more than anything else.

The next evolution for education, K-12 and above, is to adopt new standards allowing the organization to choose their modality but maintain the standard of communication.

Educational institutions need to build a data lake, or data repository, using data from all their vendors. Any vendor that cannot meet a few basic standards, needs to be eliminated from the pool of options. These standards would be simple, and would include:

  • Data, all data, can be exported in a single data pump when required
  • Data, all data, can be exported into at least one or all of the following formats: csv, tab, sql, or xlsx
  • Downloaded Data, all downloaded data, will only have encryption if the client chooses
  • APIs and other methods to sync real time data are optional; even if these tools exist, the data export requirements must be maintained

By insisting these standards be met by vendors, educational organizations will not only be able to constantly analyze all their data, they will be able to recreate themselves when they choose. Vendors will not hold the fear of data loss, or opportunity cost, over the decision makers.

The only remaining conundrum is: how do we show every school how to do this, and how to find opportunity within this new environment?

Run Self Hosted Tech Without Your School Building

Tony DePratoLeave a comment

By: Tony DePrato | Follow me on Twitter @tdeprato

Imagine you have had to evacuate your high school. Could you facilitate your classes and business processes without physically being in your building?

If most of your infrastructure is  in a cloud based environment, odds are you can maintain business continuity without your building. You might need an office or some type of staging area, but your organization can still meet it’s core requirements.

But what if you are self-hosted? What if most of your systems rely on infrastructure and data that is on premises?

This is an important conversation senior management and all the creative thinkers in your organization need to have. Here are some ideas to help guide you through the process and make (or test) your plan.

Choose a Secondary Location

Before anything technical happens, choose a secondary location to run your operations. Assume that the current location, and immediate area around the current location, are off-limits. Where can you affordably establish and maintain an operational space?

The space would need to include:

  1. Enough space for the core team to work
  2. Basic communication resources
  3. Hardcopies of data that is required to contact parents, students, and teachers
  4. Hardcopies of schedules and other data that are needed for basic school operations
  5. Basic supplies and consumables (even food and water) for the team to work for at least 14 days

Remember, this requires some minimalism. People could rotate in and out, but the core team should be as small as possible. Anyone who can work from home, should work from home.

If you can maintain business continuity without technology, then by all means try to achieve this. Every new requirement will only add complexity to the situation.

If you need technology, keep reading.

Create a Portable Network

For any data to flow inside or outside of the secondary location, a computer network is required. Here are some core items that would be required in a normal metropolitan/urban/suburban area where the mobile service has not been disrupted:

  1. One 4-5G router (if these are not available, then two or three 4-5G hotspots/wifi eggs)
  2. One high speed router to allow network and wifi customization
  3. 5-10 individual LAN cables; 1-3 Meters each; longer cables look messy but add flexibility
  4. 5-10 power extensions with sockets; avoid cheap ones
  5. Cable ties, double sided-tape, duct tape, and a few box knives

This setup will connect to the internet, and allow the small group of users to get online.

The environment will most likely be small, so maintaining basic safety when rigging equipment is essential. Damaged equipment will be difficult to manage during any type of emergency. Preventing trip wires prevents damage.

Create Portable Data

Many organizations have offsite servers that mirror their data. These organizations can keep operating without their normal infrastructure.

Most schools do not do this. Most schools cannot afford to do this. If the school is using an online classroom environment, then maintaining classes will be fairly simply as long as there is a protocol to follow. For all academics functions, creating a protocol. This should include:

  1. Attendance for teachers and students (time stamped, and strictly followed)
  2. Mimic the course schedule; when a teacher should be in math class, they should be online answering questions about math; they need to follow their schedule
  3. Establish office hours and lunch to provide some break time and organization
  4. Assign administrators to contact teachers for daily feedback and summaries
  5. Assign administrators to randomly contact students for daily feedback and summaries
  6. Send parents status updates on the situation at the same time everyday, unless a critical time sensitive issue presents itself

Using an online classroom system for your day-to-day teaching is not a requirement to have an online classroom system. A school can setup a Google Classroom or Microsoft Classroom environment for emergencies. These are usually free with educational licensing. These classroom environments need to be kept up-to-date with enrollment and scheduling.

There are a few other ways to run online classes without these cloud services, the cost is higher, but it is totally feasible. If you need this type of information, please email me: tony.deprato@gmail.com

Data files, such as spreadsheets and text documents, will be required for business functions. You may have an emergency where going into the campus is not prohibited. Retrieving hardware will be difficult and the outcome uncertain. Relying on external drives is not a great idea unless a set of those drives is stored off campus on a regular schedule.

Offsite storage is easy to manage using systems like Resilio Sync. The assumption is that the school does not want to use any cloud services. Using a peer-to-peer system would send copies daily from one location directly to the next.

There are other ways to sync files from one private location to another. Feel free to email and inquire.

Not Complete But Enough to Get Started

There are many options when designing these plans. Even if you feel the ideas here are not feasible, the questions raised are worth answering. The thought exercise should help develop policies and procedures for all stakeholders.

Administrators should be engaging their IT teams to find out how data is saved, where it is saved, and how it can be accessed. IT teams should be engaging administrators to determine the minimum core requirements for maintaining business continuity, and the total amount of downtime the organization can withstand.

 

Episode 133 – Interview Tips

Patrick CauleyLeave a comment

133

Yet another awesome episode of IT Babble. This week Tony and Patrick talk about interviewing for international schools, hypocrisy with data, Prodigy a math game and the fact that Skype is free.

Check out the talking points below.

1) Interviewing overseas and some interesting questions

  • Take many interviews
  • Having job offers help
  • Gets rid of extraneous questions
  • Common questions
  • What is your vision of technology?
  • How do you deal with difficult people?

2) Data driven hypocrisy

3) Prodigy online math game – https://www.prodigygame.com/

4) No account for Skype anymore – https://blogs.skype.com/news/2016/11/14/now-anyone-can-use-skype-right-away/

Download the episode HERE!

Subscribe to us on iTunes or Podomatic.

How Dirty is Your Data?

Tony DePratoLeave a comment

Data-Cleaning

By Tony DePrato | Follow me on Twitter @tdeprato

My basic rule for data is, unless there is a life and death scenario unfolding, bad or unclean data is not going to be used. I have yet to encounter a situation where releasing data, which will eventually wreak havoc throughout the school, is an essential and lifesaving endeavor. Delaying systems access due to data issues is difficult. Even the smallest of systems have vocal advocates who will passionately state the damage being done to learning for every day a system is offline.

The best way to exist in a data-driven environment is to be prepared. Being prepared means being aware. Awareness comes from a regular, I would argue monthly, check of all core databases and having policies and procedures for correcting problems.

The real question is this: how does someone not involved in direct data management, check data? And how does someone who is an end user of data set policies to protect the data they need?

Read More…

Some people have 10 years of experience. Other people have 1 year of experience 10 times.

Tony DePratoLeave a comment

Some people have 10 years of experience. Other people have 1 year of experience 10 times, I wish this were my quote. However, it came from a source on Slashdot – which I always recommend everyone read a few times a week.

There was an article titled, Lessons From a Decade of IT Failures :The takeaways from tracking the big IT debacles of the last 10 years. The quote actually came from the comments about the article.

It struck me in a profound way. I immediately, and sadly, thought of many of my co-workers who fall into this category. I also thought of key institutional indicators which could be warning signs that decisions are not being made from the “right place”.

The Right Place

 

There are many schools that run teacher centered, adminstration centered, and community centered models of education. These can all be reviewed at another time, but what they all have in common is that the needs of the student are not the priority.

Research in Education in the last twenty years overwhelmingly supports student centered learning. To be in the right place a school should be following student centered approaches. This requires fairly frequent adjustments to scheduling, assessment practice, learning support, etc. Being student centered means supporting a culture of change. Not always large swooping change, but often small adjustments that ripple influence like a stone hitting the water.

Key Indicators of a Problem

If change is supported in a student centered environment, a school administrator would not see the following (would not, think negative, think dark):

  • The same schedule being followed for more than three years
  • No curriculum revision cycle
  • Lack of data for moving students to different levels
  • A small number of PD requests from teachers
  • The absence of a formal school improvement plan
  • No effort given to defining of hours related to subject completion or academic success
  • People in non-leadership roles running programs from a “playbook”
  • Teachers and Administrators without improvement indicators attached to their annual review

Technology Can Help

Of all the things we can use technology for in school, nothing is easier and more clear cut than using it to collect and study data. From basic Excel implementations to Powerschool, there are many options to allow a small group of administrators to collect and study data.

This process, and hopefully a regular one at that, would quickly flag trends leading to the negative list of key indicators above.

Finding the problem after it has occurred is not going to be enough. The only way to have a real solution, is to stop the problem before it reaches a critical mass and becomes embedded in the culture.

Like a video game with flaw or loophole: If you detect it before you launch the game then it is classified as an error; if you detect it after you launch the game it is classified as a feature.

Tony DePrato

http://www.tonydeprato.com

http://www.intslead.com/

 

Research First, Angry Email Later

Tony DePratoLeave a comment

pointing-finger-xcg67bAcA

Despite the title, I am not in favor of sending angry emails. I use to be known for my emails, sent out of frustration, and also justified in content. However, in hindsight, I regret all of them that could have been communicated face-to-face, or after the frustration had faded.

I received an email such as this, this week. I received it after 9:oo PM, and late night bad emails are the worst. I attempt to not read email after 7 PM, because I know in an emergency people will call me, and I know late night emails are often full of stress and regret.

After reading the email, I did have to reply. I knew this person might start being aggressive with my staff the next morning, and I would not be there based-on my work plan (I cover two campuses in different areas of my city). So I asked that they give me time to research the problem, and reminded them that even though the fault seemed to fall on the IT support people, in fact, those same people had been dependable in the past. And that those people also have been working overtime to accommodate jobs that are not part of their employment contract. These are people who are mindful enough to call me when they know they have made a mistake or forgotten something.

Do Your Research

There is data everywhere. School administrators and heads of department need to start paying attention to where data comes from and where it is stored. Card swipe systems, cctv footage, security guards walking around(with smartkeys), email time-stamps, network access logs (when someone logs in and out), etc. The first step in any “he said she said” should be accessing data and not the anecdotal kind.

I always take time to look at the environment, the people, and the systems surrounding them to determine if an impartial data source exists.

Separate and Report

Often when people collect data, they get the pieces they need and make a report. The people reading the report do not see all the data (Sounds like a congressional trial transcript to me).

The best practice is to make a copy of the raw data, and archive it. I think a zip archive is an excellent way to do this, as it creates metadata. It will be clear if there was any manipulation in the data file used in the report.

Then, instead of deleting, highlight the data points that are valid for the argument.
It is ok to sort them to the top of a list, but try not to remove the other pieces. If you argument is sound, and/or, the best of the worst based on what is known (also very congressional), then the extra data is not going to hurt the report. In fact, someone may find something useful that was missed.

Then make the report, and do not use the words feel, believe, or think. State the facts that the data supports first. Then add verbal accounts of the event. Using emails as data works best if you print them as PDF files and then highlight the time-stamp and key details. Do not copy and paste emails, even forwarded emails can seem bogative.

How Did It End?

My department was not at fault. The event in question ran outside of the contractual working hours, so the space was prepared by the IT support before they left. Their two points of contact failed to respond to communication, but the room was ready. However. the event organizers did not inform security. Security walked into a large room, saw that everything was left on, and in a moment of global awareness, they shut-off the power.

After pulling the data from the door swipe system, I found that security went into the room 20 minutes after the IT support people. This room is the only place power can be accessed.

A simple call to the person overseeing the building revealed no event has been logged for that evening, so basically, no one knew what was happening. Security wanted to save the school a few dollars in energy bills, and that was the end of the story.

I have taken the opportunity to allow the person who sent the email to ponder the data.
I am hoping they do the correct thing and inform everyone what really happened, as I suspect I was not the only email recipient that evening.

Right or Wrong, Let It Go

The wrong move in situations like this is to alert everyone that you are, in fact, awesome.
The wrong move is to make everyone angry when you are correct, ensuring that they skewer you the minute you are wrong. You will be wrong. Your team will make mistakes. You will make a plan that ruins someone’s day. This is going to happen. If you want forgiveness, be prepared to be forgiving, and give people a chance to correct their mistakes.

Data is like The Force, it cannot be welded in the presence of anger or it will turn you to the Darkside.

Tony DePrato

www.tonydeprato.com

How a ‘C-‘ Can Be a Good Thing

Tony DePrato2 Comments

It’s about getting things down to one number. Using the stats the way we read them, we’ll find value in players that no one else can see.~ Moneyball 

I often take ideas and assign a numerical scale to them, in order to compare them to other things. I always tell people that they can “make their own math”. Most people just stare at me blankly, and others just laugh like I am joking.

Believe me, all that math you may have partially studied in school is useful. Somewhere along the line someone forgot to add a few key points to the math textbooks. For example emphasising  that if a bunch of people sitting in a room can decide if a movie is a G, PG, PG-13, etc., then any group of people sitting in any room can do the same thing.

Even though they do not realize it, educators do this all the time when they make decisions about grading, grading scales, and standards. Recently I have been looking at grading scales for a Shanghai Primary School, a Shanghai Middle School, a year 9-10 IGCSE program, and a year 11-12 IB program. In my current position I am involved in implementing these scales among a common population of students.

These students will start on one scale and finish on another. They will go from letters, to numbers, to different letters, and back to numbers.

It is perplexing when considering the transcript and the key needed to decode the transcript.

I think the way schools report progress is a bit insane. It seems logical to give a student a number or letter and say, “This = Good and That = Bad” . However, over the course of time, the standards connected to these metrics change. So the logic does not hold up.

Trying to report the standards seems logical, but the number of standards per student, per subject, and per grade overtime would be overwhelming for most people to read and interpret.

So how should schools get things down to one number, using the information the way parents and students need to read it?

I suggest the answer is to stop reporting numbers and letters, and to start reporting trend lines. 

trending

Trend lines not only show a student’s performance  overtime, they clearly show if the student is on a steady, moderate, or rapid incline or decline. A trend line can group categories of things into single points, and those points can be reviewed quickly. Any points of concern can be expanded for conversation.

The most interesting thing is that a student who previously had an ‘F’ in science, but now has a ‘C-‘, will appear to be improving (Which is good, because they are improving). A grade of 55 that is now a 71 shows a 30% improvement. If this was a mutual fund, you would be smiling.

Currently, what do parents and students see in this situation? They see an ‘F’, and a ‘C-‘.
That does not seem like much of an improvement when the grade is explained as below average and described as needing significant improvement.  An 81 changing to a 91 looks great, but that is just a 10% improvement.

The truth is, the trend line would show not only improvement but some degree of effort. Effort that is not calculated by someone’s opinion, but through the interpretation of data.

Tony DePrato

www.tonydeprato.com