AI Ethics

A growing area reflecting the impact of technology on society is ethics and AI.  This has a few variations… one is what is ethical in terms of developing or applying AI, the second is what is ethical for AI’s.  (Presumably for an AI to select an ethical vs unethical course of action either it must be programmed that way, or it must learn what is ethical as part of it’s education/awareness.)

Folks playing in the AI Ethics domain include a recent consortia of industry players (IBM, Google, Facebook, Amazon and Microsoft), the IEEE Standards folks, and the White House (with a recent white paper).

This is a great opportunity for learning about the issues in the classroom, to develop deep background for policy and press folks — concerns will emerge here — consider self driving cars, robots in warfare or police work, etc.  and of course the general public where misconceptions and misinformation are likely.  We see many movies where evil technology is a key plot device, and get many marketing messages on the advantages of progress.  Long term challenges for informed evolution in this area will require less simplistic perspectives of the opportunities and risks.

There is a one day event in Brussels, Nov. 15, 2016 that will provide a current view on some of the issues, and discussions.

 

Health App Standards Needed

Guest Blog from: John Torous MD, Harvard

Last year, the British National Health Service (NHS) thought it was showing the world how healthcare systems can best utilize smartphone apps – but instead provided a catastrophic example of a failure to consider the social implications of technology. The demise of the NHS ‘App Library’ now serves as a warning of the perils of neglecting the technical aspects of mobile healthcare solutions – and serves as a call for the greater involvement of IEEE members at the this evolving intersection of healthcare and technology.

The NHS App Library offered a tool where patients could look online to find safe, secure, and effective smartphone apps to assist with their medical conditions. From major depressive disorder to diabetes, app developers submitted apps that were screened, reviewed, and evaluated by the NHS before being either approved or rejected for inclusion in the App Library. Millions of patients came to trust the App Library as a source for high quality and secure apps. Until one day in October 2015 the App Library was gone. Researchers had uncovered serious privacy and security vulnerabilities, with these approved apps actually leaving patient data unprotected and exposed. Further data highlighting that many approved apps also lacked any clinical evidence added to the damage. Overnight the NHS quietly removed the website (http://www.nhs.uk/pages/healthappslibrary.aspx) although the national press caught on and there was a public outcry.

As an IEEE member and a MD, I see both the potential and peril of mobile technologies like apps for healthcare. Mobile technologies like smartphone apps offer the promise of connecting millions of patients to immediate care, revolutionizing how we collect real time symptom data, and in many cases offering on the go and live health monitoring and support. But mobile technologies also offer serious security vulnerabilities, leaving sensitive patient medical information potentially in the public sphere. And without standards to guide development, the world of medical apps has become a chaotic and treacherous space. Simply go to Apple or Android app stores and type in ‘depression’ and observe what that search returns. A sea of snake oils, apps that have no security or data standards as well as no clinical evidence are being marketed directly to those who are ill.

The situation is especially concerning for mental illnesses. Many mental illnesses may be thought of in part as behavioral disorders and mobile technologies like smartphones have the potential to objectively record these behavioral symptoms. Smartphones also have to potential to offer real time interventions via various forms of e-therapy. Thus mobile technology holds the potential to transform how we diagnose, monitor, and even treat mental illnesses. But mental health data is also some of the most sensitive healthcare data that can quickly ruin lives if improperly disclosed or released. And the clinical evidence for the efficacy of smartphone apps for mental illness is still nascent. Yet this has not held back a sea of commercial apps that are today directly available for download and directly marketed to those whose illness may at times impair clear thinking and optimal decision making.

If there is one area where the societal and social implications of technology are actively in motion and needing guidance, mobile technology for mental healthcare is it. There is an immediate need for education and standards regarding consumer facing mobile collection, transmission, and storage of healthcare data. There is also a broader need for tools to standardize healthcare apps so that data is more unified and there is greater interoperability. Apple and Android each have their own healthcare app / device standards via Apple’s ReseachKit and Android’s Research Stalk – but there is a need for more fundamental standards. For mobile mental health to reach its promised potential of transforming healthcare, it first needs an internal transformation. A transformation led in part by the IEEE Society on Social Implications of Technology, global mental health campaigns (changedirections.org), forward thinking engineers, dedicated clinicians, and of course diverse patients.

If you are interested in tracking standards and developments in this area, please join the LinkedIn Mobile Mental Health App Standards group at: http://is.gd/MHealthAppGroup


 

John Torous MD is an IEEE member and currently a clinical fellow in psychiatry at Harvard Medical School. He has a BS in electrical engineering and computer sciences from UC Berkeley and medical degree from UC San Diego. He serves as editor-in-chief for the leading academic journal on technology and mental health, JMIR Mental Health (http://mental.jmir.org/), currently leads the American Psychiatric Association’s task force on the evaluation of commercial smartphone apps, co-chairs the Massachusetts Psychiatric Society’s Health Information Technology Committee.

Smart Government: ICT Enabled Social Engagement in Public Organizations

An SSIT Guest Blog provided by: Carlos E. Jiménez; Open & Smart Gov Specialist, IEEE SSIT Sr. Member; Barcelona, Spain.

In a broad sense, we usually use e-Government concept as the ICT adoption by public organizations as helpful tool in order to improve the way they achieve their goals. Key elements in these organizations are elements like efficiency, effectiveness, transparency and citizen-centric oriented.

However, it is important to say that in a more specific sense, there are important differences when we talk about its degrees and elements within this field. Then, we could talk on 4 distinct concepts: e-Administration, e-Government (in a more specific sense), Open Government and Smart Government. These stages are incremental where ICT transform the public organizations at the same time as they produce better services to citizens.

In the table, we can see that e-Administration started with the ICT adoption addressed to automatize workflows in public organizations (1st stage, -Bureaucratic organization) and, later, the e-Government stage (2nd stage, -Professional organization) includes interaction between citizens through the use of electronic tools, as well as bi-directional flows of information allowing citizens to use e-services. Next, technologies contribute and facilitate the move to a 3rd stage (Relational organization) where -Open Government- is achieved, allowing a high degree of the governance paradigm and not only through the use of e-services. In this stage there is a participation of the society in decisions and processes that before, were mainly done exclusively by the organization. A 4th stage and type of public organization (Intelligent organization) after the Relational one, would be based in the optimized IT adoption degree, and how it can transform the public organization as well as society.

Organization  Modernization Level ICT Role
1. Bureaucratic Begin Automatized Workflows  (e-Administration)
Benefit: increased internal efficency
2. Professional Middle Citizenship Interaction (e-Government).
Benefit: efficient public services (filing forms…)
3. Relational Advanced Citizenship participating in governance (Open Government).Benefit: Paradigm of governance
4. Intelligent Optimal:
Adopted completely Interoperability principle and Open Innovation as tool
Interconnected Ecosystem (Smart Government)Benefits: real time, data driven – integration of information, Public-Private-People Partnership…

This 4th “refined” public organization level, would be achieved as a result of ICT as tool that is being used in perfect harmony with: a) Open Government b) the Social & Open Innovation in public organizations and c) a maximized Interoperability Principle [this concept is expanded in a special issue of IEEE Computer Magazine, Oct 2014]. The concept of Smart Government, then, will have all these factors, and the social implications of technology are being key here.

Indeed, we have to understand that territories and cities only will be smarter if and only if are more social, through thinking in the best options for their citizens, specially, avoiding negative impacts of technology. To get a sense for how this looks in practice see, in the case of Barcelona, https://smartcitizen.me/.

What areas of government in your territory are starting to move towards the “Smart Government” level?

 

Hard Hitting Robots (Not) — and Standards

ISO is developing standards for the contact between collaborative robots and humans working in close proximity.  Which raises a question of how hard a robot can hit you, legally. Of course this also raises concerns in industry about liability, work-place-safety legislation etc.

There is nothing new here in reality.  Humans have been working in collaborative environments with machines, animals and even other humans. In the U.S. some of the first workplace limitations were actually triggered by animal cruelty legislation applied to child labor. And of course as our experience has increased with different types of equipment, so have the sophistication of work-place protections.   Industry leaders working in these areas should be pro-active in helping to set standards, both to have a voice in the process, and to protect workers.  Well considered standards provide protection for employers as well as workers.  Of course when insufficient diversity of perspectives establish the standards, they can end up with an imbalance.

In my own experience, which involved significant work with standards (POSIX at IEEE and ISO levels) industry is wise to invest in getting a good set of standards in place, and users/consumers are often under-represented.  IEEE and IETF are two places where technologists can participate as individuals, which provides a better diversity in my experience. ISO operates as “countries”, with some countries under the lead of corporate interests, others academic, some government agencies. In general I suspect we get better standards out of the diversity possible from forums like IEEE — but with industry hesitant to fund participation by individual technologists, these forums may lack sufficient resources.

One of these days, you may get a pat on the back from a robot near you. Perhaps even for work well done in the standards process.

Mr .Jim Isaak


Mr.Jim Isaak
Bedford, NH; United States
SSIT Volunteer since: 2.003
picture of Mr.Isaak SSIT Roles
Vice President, 2015; Blog master 2014-present
IEEE Roles
SSIT Board (elected or appointed), IEEE Director, Member IEEE Technical Activities Board, IEEE Standards Association, IEEE Section/Chapter
SSIT 5 Pillars Interest:
Sustainability, Ethics, Impact of Emerging Technology
Other:
privacy, predictive (science) fiction, policy
Web site & Social Media
http://JimIsaak.com
http://www.linkedin.com/in/Jim-Isaak
IEEE Senior Member in New Hampshire Section of Region 1

Last updated: 29/01/2017