Designing Agentive Technology

by Christopher Noessel

Advances in narrow artificial intelligence make possible agentive systems that do things directly for their users (like, say, an automatic pet feeder). They deliver on the promise of user-centered design, but present fresh challenges in understanding their unique promises and pitfalls. Designing Agentive Technology provides both a conceptual grounding and practical advice to unlock agentive technology’s massive potential.

Genres: Design, Technology, Artificial Intelligence, Nonfiction

280 pages, Kindle Edition

Published May 1, 2017

Reviews

[Megan Rivera · Rating 4 out of 5]

This was a very informative book about Agentive technology. I was able to learn about the Drebbel’s Incubator which, was used to regulate temperatures. It is important to lower the amount of work when it comes to temperature usage. It was great to see the methods that work and didn’t work to help conserve energy. I enjoyed reading this book.

[Find me on The Storygraph · Rating 3 out of 5]

Ik vind het lastig om een recensie over dit boek te schrijven. Het is een aardige introductie tot het fascinerende veld van kunstmatige intelligentie (AI), geeft veel goede en nuttige voorbeelden van AI in de praktijk van vandaag en de toekomst en schetst diverse ethische en praktische bezwaren. Maar het doel van het boek (of was het mijn verwachting?) was een begin maken met nadenken over Interactiepatronen voor AI. Het lastige is natuurlijk dat dit een volledig nieuw veld is. Over vijf, tien jaar gniffelen we om de onzekere zoekende toon van dit boek. Tegen die tijd zijn de patronen uitgekristalliseerd. Nu weten we alleen nog maar dat we niets weten.
Als je een introductie zoekt tot AI, misschien nog even verder kijken naar een beter boek. Maar Noessel is een onderhoudend schrijver en een autoriteit als het gaat om interactiepatronen. Met andere woorden: ik weet het niet...

[Scott Pearson · Rating 5 out of 5]

It seems that advances in Artificial Intelligence have always been just over the world’s horizon ever since the 1950s. Although General Artificial Intelligence (AI) – the kind that mimics full human intelligence – still seems decades away, our society seems to be adapting so-called Narrow AI already. (In fact, for work, I develop software that uses Narrow AI approaches for use by research administrators.) Noessel seeks to explain how non-programmers can design software that helps us humans do our lives better.

This book helps everyday people identify patterns in life that can be ameliorated through the use of technology. It’s not geared towards technical people like software developers or engineers as much as towards the general public. Ultimately, the public is going to be designing how this new software works, and these conversations need to be had at a societal level. Noessel uses plain examples and even explores a case-study of someone building an AI-equipped home garden.

In coining a new term, he calls this type of technology “agentive.” These are not just assistants to human tasks but “agents” that go out and do things for us. Technologists, like myself, are already developing software in this niche, and popular software companies – like Google, Apple, or Microsoft – are already fleshing out what these models mean in their products. Through this work, Noessel seeks to help us understand how we can adapt these advances to maximize beneficial human impact. The best technology relies on human reflection so that humans remain in the proverbial driver’s seat.

Of course, much discussion about the ethics of AI needs to happen as well. Noessel’s penultimate chapter, perhaps the longest in the book, focuses on ethics. Will this new paradigm lead to a utopia or a dystopia or something in-between? Much ink (or e-ink) has been spilt over generalized AI (can robots be human?), but not enough has been spilt over how to live in a world where most things are out of direct human control. Will the stock markets successfully adapt to algorithmic trading? What of the future of war? Will important human niches of knowledge be forgotten? Can people choose not to work in this new world? These questions are immensely interesting and need to be discussed before the situations present themselves.

Overall, this book addresses many real dilemmas faced by the march of technology. I like to keep up with software design to supplement and empathize with the designers who are my colleagues. It provided a dispassionate yet interesting exploration of things software deals with each day. It also brings to light many of the issues that I believe the wider public should deal with. Thinking about these things will facilitate their successful adoption by humans. Many in the software world are concerned that technology not become its worst impeding enemy; the reflections in this book advance that case, and for that reason, I’m grateful I stumbled across this book.

[Ella · Rating 5 out of 5]

I am very impressed with this book. Not only does it do a great job at describing what is Agentive Technology (a relatively new term even for me), but it does so in such a structural way that you get to understand how Humans slowly changed their "agents" throughout history (with the example of the thermostat) until having reached software and created AT. It goes trough very important topics such as: How do we identify what is an AT, what is AT build for, how is it supposed to work with different use cases, how will it interact with the users, what is it supposed outcome?

It explains everything so very well. The book knows how to entertain you, how to explain concepts that you may not be familiar with, provides examples for you to follow along. You get to grasp how the design of software that is AT is built, is tested and then launched.

The book, of course, also addresses a very important question when it comes to AI. How will this affect us, good and bad? How dependent will we become of it? Will we use our skills because of their assistance? Will they replace us?

I learned so much from this book, and it would totally be a book I recommend to anyone who is interested in AI, specially in Agentive Technology when it comes to designing or just learning of it. Definitely worth a read. It is a magnificent book.

[sarah semark · Rating 4 out of 5]

This isn't quite what I'd expected: I think I was expecting a more general approach to AI, but instead it's geared specifically toward what the author refers to as agentive technology—technology that acts on the users' behalf (like a Nest thermostat or a Google alert) after being set up with some initial preferences or instructions, then monitoring for trigger events, and acting according to the user's instructions.

Nonetheless, it's a super interesting and accessible look at how this technology is currently being used, and where it may lead in the future. Worth reading even if you aren't building something agentive right now, since there are some good insights about AI generally, and it'll encourage you to think a bit more agentively about the things you're currently building as well.

[Het · Rating 4 out of 5]

Fascinating read.
The author has written a well planned book wherein initially the basic is explained before going into the depths. Even then it is divided into sub questions of why, what and how. There are lots of examples with the simple language makes it easier to understand. It was informative reading about AI. What makes it a wholesome package is the inclusion of ethics towards the end mainly for understanding the misuse along with the use.
So even if your are not tech savvy, your imagination can make it a wholesome read.

[Shankar Balasubramanian · Rating 3 out of 5]

This is a really good book, especially the first two parts. Chris gives us all some fine frameworks to design with AI and provides great examples. The last part of the book somehow doesn’t feel as tight, delving too much into philosophy and ethics. It is necessary, but could be shorter with that part. The book has a recap section after every chapter which makes it easy to refer later. If you are in tech, this book is a must read. I wish it was better in the last part though.

[Gabriela Véghová · Rating 4 out of 5]

The book is great in describing what is an agentive technology (a technology that does the job for you). It compares it to assistive technology (technology that helps you do the job) and it mentions their advantages and limits. Also great in describing the principles for designing a successful agent. An experienced designer would probably figure it all out - sooner or later. However, having them well described in one place is really helpful. In the beginning, it might seem that agentive technology can be anything and everything, but it gets much clearer throughout the book. The last part didn't resonate too much with me. It discusses philosophical and ethical questions regarding agentive technology in a very broad sense. I would much more appreciate discussing everyday problems of agents, like security and privacy. I'd be interested in real examples and possible solutions.

If you're interested in AI, better go for a different book. Here it's only mentioned as an enabler and actually, you don't really need any AI for most agents. Also, most of the examples don't actually work with AI.

[Richard · Rating 5 out of 5]

This is one of the few books about designing for AI and a great place to start. I've seen the author at conferences before and I really appreciate the way he explores and approaches the intersections of design and technology. The book is an easy read with great examples, a running case study, and end of chapter summaries. I wish more design skills books were like this.

[Mike Hales · Rating 5 out of 5]

Glorious mixture of thought-provoking points and real world examples. Read this book as I started my journey into building ai products and stimulated some great ideas as well as valuable warning signs or caution points.
Highly recommended.

[Jon Daiello · Rating 3 out of 5]

Overall this is a great introduction for designers and product leaders to understand the concept of creating agents that work for the humans they build for. Lots of good frameworks to use and basic guidelines to ensure you ship valuable experiences and not just a party trick with tech.

[Peter · Rating 3 out of 5]

Not that many books about AI design yet, this one summarizes some of the context.