Ideas Blog

AI Update: Have you seen my Putin Impersonation? It’s a blast.

URKDagg

Two recent articles in the media reminded me of a concern about AI.

Firstly, the BBC asks us whether we would care if a feature article was, ‘written by a robot’? The implication is clear, it will soon be the norm for digital content to be created by intelligent systems.

Secondly, and more menacingly, another BBC report suggests that hacked video and voice synthesis tools will soon be producing lifelike quality. The actual report cited by the BBC story can be found here: The Malicious Use of AI Report.

Impersonating humans

It may not be long before digital tools can convincingly (read indistinguishably) produce fake video and speech, thereby impersonating specific human beings. This content could be augmented with mannerisms and linguistic style harvested from previous online posts, speeches, comments or video produced by the target.

What emerges is a simulacrum. Reality recedes and we can no longer tell what is real and what is not.

The problem with all this is that very soon the capacity of someone, or some system with advanced voice, video, and linguistic style tools at their disposal will be able to convincingly impersonate human beings. To all intents and purposes they will be able to hack reality.

Impersonation does not mean that the system doing the impersonating will need to interact with people or pass a Turing test. All that is required is that the content produced is convincing.

It is not completely in the realm of fantasy to picture ‘Vladimir Putin’ giving the order to use nuclear weapons, and this order is either acted upon, or retaliated against. Many other undesirable situations are possible, both mundane and terrifying.

A host of companies are already working very hard, and with much success, to create impersonations, of their own staff, for the purpose of interacting with customers.

Currently many states have laws forbidding the impersonation of a person, by another person, but do our laws adequately forbid the impersonation of a human being by a digital system? And even if it forbidden, how will we enforce this?

Trust: the next big human problem

Humans over time have struggled with some key societal problems, and found solutions. These include:

  • the problem of coordination and cooperation, solved by language and written symbols
  • the problem of exchange, solved by money
  • the problem of information dissemination, solved by the printing press
  • the problem of scale, solved by the industrial revolution

We now must solve The Problem of Trust and Authenticity.

This has been demonstrated by the debacle around Fake News, where Facebook and other digital media companies are scrambling to implement change.

We need to solve the issue of trust, at the levels of moral norms, digital systems, societal systems, and laws.

Technologies like blockchain are small steps in the right direction, but the problem persists and other solutions and constraints are needed.

We need to have a conversation as a society about what kind of future we want to live in, and what limits, laws and norms we want to impose on emerging technologies, particularly artificial intelligence, which by it’s very nature mimics us.

This conversation needs to begin now and it needs to involve the technology sector, a fully informed general public, and the government.

Click here to listen to a talk by Adapt Research Ltd’s Matt Boyd on AI and the media given at the NZ Philosophy Conference in 2017. 

From Russia with Love: Time for serious work on the benefits and risks of artificial intelligence

The following introduction is an excerpt from a just published blog at Sciblogs that examines the role of AI in society and democracy. Read the full article here

Transformative advances in artificial intelligence (AI) have generated much hype and a burst of dialogue in New Zealand. Past technological change has led to adaptation, but adaptation takes time and the pace with which AI is arriving appears to be accelerating. For example, recent news about the unfolding ‘Russia Investigation’ may be just a prelude to what is possible if AI tools hijack our social systems.

Technology offers us opportunities to do things we previously could not, but in doing so the use of technology also changes us, and it changes the systems and norms of society.

Read more…

Artificial Intelligence, Freedom and Democracy: Talk at NZ Association of Philosophers Conference Dec 4, 2017

“Artificial Intelligence and Free Will: A 2017 Christmas Carol”

(Talk presented by Matt Boyd at the NZAP Conference)

 

Technological innovation builds our minds, does it build society too?

 

screen-shot-2013-07-23-at-3-35-03-pm

As we build our world we build our minds: reboot

Six years ago, while I was writing my PhD on technology and human nature, I wrote a blog where I argued that:

  1. Context builds us – Our social and technological environments can hinder, but they can also drive psychological development.
  2. Technology drives human development – Physical and digital tools shape us and build our intelligence; small doses of technology cause transient changes and long-term exposure has lasting effects.
  3. Our minds depend on technology – Much of what is unique and modern about human minds depends on technology for its development.
  4. The technology we build, in turn builds us – Basically, as we invent stuff, we change the context of development for the next generation, and they grow up with different thinking and affordances.
  5. Hence, technological innovation causes and sustains psychological evolution.

This reasoning was mostly based on the interplay of relatively static technological tools and human minds. However, we now see a range of dynamic and powerful technological tools emerging. This has important implications for human nature and the nature of society.

Building our minds

The symbols, media, tools, and methods that we invent shape and extend our minds. This is our brain’s ancient trick and amplifies what we are capable of achieving. We frequently make use of external supports to extend our brain’s capabilities (think words, lists, numbers, the abacus, abstract diagrams, calculators, number lines, and a range of other tools).

On the other hand, without technology we are mentally crippled. Without our abacuses and iPhones we are like Alzheimer’s patients without their post-it notes and pill schedules. As well as appearing rapidly, a lot of recent human evolutionary psychological advances could disappear overnight should the technological context sustaining them change.

These changes can be slow or fast. Geological processes are usually slow and ancient, but earthquakes show us that they can be sudden and dramatic. In analogous ways, small changes to technology can change the cognition of a population slowly over time, but significant innovations, such as mathematical symbols or the internet, can have sudden unexpected effects. With new technologies come new implications. Cue artificial intelligence…

The reason we see these effects is because the brain is a highly malleable organ. Stimulated at the right stage of development it can be made to do, or not do, almost anything.

But the more we offload responsibility and dynamic cognitive processes to intelligences other than our own, the more we risk becoming automated masters of our own creations. Instead of technology augmenting our intelligence, we risk merely obeying algorithms. If thought is offloaded to digital supports and never re-internalized, the cognitive loop is broken and instead we divest cognition, and therefore power and control.

A mother’s diet has a critical effect on the future health of the unborn. In similar fashion technological diet in childhood shapes thought processes. Given our ability to build a range of different technological environments for our children, then it is likely that our innovation wittingly or unwittingly causes an array of emotional and psychological traits. Ever since technology was invented, as we build our world we have been building our minds.

Building society’s future

That was my conclusion in 2011, but there is an important extension of the argument to society and democracy, as noted in a Scientific American article this year.

Society and social structures are just as malleable as human minds. The technological environment of a society produces a set of affordances, and with affordances come possible actions, institutions and norms. The technological environment, coupled with our tendency to now offload dynamic processes (once the domain of thought) to digital systems, means that technological uptake is wittingly, or unwittingly building the nature of our society.

The old dogma was that technology does not determine people, because it is how we use technology that is important. However, in the new world of dynamic and autonomous technologies that dogma must be called into question.

The moral of the story is that we must reflect carefully on the possible consequences of introducing even benign seeming technologies, and uphold a principle of precaution and willingness to respond with rules and norms should things not turn out as we expected.

At Adapt Research Ltd we are very interested in the social, philosophical, psychological and ethical aspects of technology and innovation. Contact us here to continue the discussion.

PHARMAC’s ‘Factors for Consideration’, Justice, and Health Need

MedicalMalpracticeCasesAndTrial

When deciding what medications to publicly fund PHARMAC uses multiple decision criteria, one of which is ‘health need’. So how can we establish who needs what in healthcare?

Distributive Justice

One approach is to take the perspective of justice. What factors do we need to consider to ensure a just distribution of resources? John Rawls provides an answer to this question by inviting us to consider what kind of society we would want, but we must consider it from an original position, behind a ‘veil of ignorance’ where we do not know who we will be in this society, or what our circumstances will be like.

Rawls thinks we would come to two conclusions.

Firstly, we would want there to be rights. In the case of healthcare everyone would have a right to healthcare because no one knows from the original position whether they will be sick or healthy.

Secondly, the only inequality in healthcare that ought to be pursued is inequality that also raises the health of those who are worst off. An example of this might be colorectal cancer screening programs, which are shown to widen health inequalities, while making the worst off better off. Overall, the aim of resource distribution should be to maximize the health of those worst off. This is deduced logically, because from the original position we ask ourselves, ‘what if we were the worst off?’

Impact of Justice on Population Health

This means that logically a minimum level of health will emerge, this occurs because all health resources will be distributed in the first instance to those least well off, to raise their quality of life to the degree currently possible with existing treatments.

Resources will also be justly given to those better off, if the process raises the level of health of those least well off. For example, the colorectal screening program identified above, or perhaps other health resources that improve the health of those already well off so that they can better care for those less well off.

Once those least well off have been allocated benefits to raise them to the level of the next least well off, or once they have been allocated all existing reasonable treatments, then we move allocation to the next least well off, and so on.

What might PHARMAC do?

So, how ought PHARMAC to interpret ‘health need’ from this viewpoint on distributive justice? I raise five issues:

  1. PHARMAC currently considers ‘government health priorities’ – this is fair enough, provided these priorities are: (a) looming big expense items (e.g. due to demographics or epidemics), (b) aimed at addressing unjust health inequality, or (c) targeting those individuals who are living below some minimum standard of health (this is the maximizing the minimum approach favored by a Rawlsian concept of justice).
  2. PHARMAC currently considers the ‘availability and suitability of existing treatments’ – this is also fair enough. The concept of a minimum standard of health ought to be important here. From the original position, we would all want to ensure that those who are very unhealthy are supported towards health if possible, whereas we would be less concerned about increasing health of those already in reasonable, though not perfect health (their health need is lower). There are usually diminishing returns by continuing to spend on those already nearer to full health but more importantly this does not help those worst off.
  3. PHARMAC considers the ‘health need of the person’. This should be important but only in the context of the population. This is a critical qualifier. The person only has a health need if they are below the mean or minimum standard of health for the population. If they are not then they don’t have as much need, but others who are below the standard do have need.
  4. A further point when considering need, is that quality adjusted life years (QALYs), which are the unit of accounting used by PHARMAC to designate utility, are not sufficient measures of worthwhile life. An example illustrates the point. It might be very meaningful for a grandparent to stay alive until her great grandchild is born. This could be true even if this means living a year at low quality of life rather than 6 months at higher quality of life. The person may prefer the first situation even if it amounts to fewer QALYs. So again context is critical.
  5. In pursuing the logically derived minimum standard of health (deduced from an impartial original position and the health budget) then there are two important needs: (1) cure for people suffering ill health, up to the level of the next worst off, iteratively. And (2) prevention, to stop people from dropping below the minimum standard. The concept of prevention is important, and it allows for allocating resources to those who are more well off currently, because it maximizes health resources available downstream to help those least well off. Preventive need is determined by the probability and time course of dropping below the level of the least well off. Curative need is determined by the probability of success of the treatment (reasonable chances) and the magnitude of the gain (up to the point of the next least well off).

There’s a lot more to be said on distributive justice in health care as informed by a Rawlsian viewpoint. But these points are a good place to start discussion.

To vape or not to vape: is not the right question

snus

So you’re a smoker, you’ve tried to give up, many times, but can’t.

You are well aware of the many health risks, and you want to minimize them. What do you do now?

The New Zealand Ministry of Health has just published new guidance on e-cigarettes. It is still unclear how successful vaping is in helping people quit, but in my opinion that is not the issue.

If you want to quit then use one of the methods that has been proven to increase your chances in randomized controlled trials. These include various forms of nicotine replacement products (patches, gum, inhalers), or cognitive behavioural therapy.

However, if you have tried to quit several times, then minimising the damage to your health is now the priority. Vaping could be seen as a reasonable alternative to smoking for those who cannot quit. Especially since the Ministry notes that: ‘Expert opinion is that e-cigarettes are significantly less harmful than smoking tobacco but not completely harmless.’

Other nicotine products

However, e-cigarettes are only one product in a suite of nicotine containing products that include smokeless tobacco such as chewing tobacco and Swedish snus, a tobacco product placed under the lip against the gums.

Recently New Zealand has made moves towards legalising these products as well, see here for details.

So is this a good thing for the health of Kiwis, and how harmful is ‘not completely harmless’?

The answer is that these products are almost certainly safer than continuing to smoke.

Smoking tobacco has strong causative associations with a range of health conditions including: laryngeal cancer, lung cancer, oropharyngeal cancer, gastric cancer, colorectal cancer, liver cancer, pancreatic cancer, bladder cancer, coronary heart disease, hypertension, strokes, chronic obstructive lung disease, and it exacerbates influenza and other infections. And that list is just the beginning. Smoking tobacco has positive associations with well over 50 diseases.

In comparison, nicotine-only products in various forms have equivocal associations with: laryngeal cancer and lung cancer; have not been associated with bladder cancer, coronary heart disease, pancreatic cancer, esophageal cancer; and there is insufficient evidence to determine any association with obstructive airways disease, gastric cancer, and colorectal cancer. The list goes on…

Nicotine alone is actually beneficial in ulcerative colitis, and it aids cognition.

Whether you quit smoking (or nicotine), or don’t quit, isn’t really the issue. The goal really is to be as healthy as possible. So what matters most is the health outcome data for the product that you are using.

Contrary to the Ministry’s advice nicotine does appear to have some health risks. For example, snus use has been associated in large studies with hypertension in men.

But these risks need to be weighed against the risks of smoking for those who cannot quit. And the risks to others posed by environmental exposure to tobacco smoke.

Clearly further studies are needed in this space. However, when stuck between the devil of smoking and the deep blue sea of snus, the smokeless tobacco products appear to be shallow and mostly without waves. This has got to be better for society than the status quo.

Funding a new drug: A Devil’s Advocate, back-of-the-envelope approach

funding-hospital-and-healthcare-projects-in-a-difficult-market

The Guardian headline reads: “New heart treatment is biggest breakthrough since statins” and the article goes on to claim that “cancer deaths were also halved”.

Sounds impressive. So how should we decide whether to fund this drug?

The New England Journal of Medicine published the findings of a randomised controlled trial into this new treatment just two days ago.

The study was well constructed, with an impressive sounding 10,061 participants. All participants had previously had a heart attack and had high levels of one marker for inflammation.

The study counted whether participants had another heart attack, stroke, or died during the follow-up period of approximately 4 years.

During that time 16% of people taking the placebo suffered one of the primary outcomes (535/3344) compared with 14% (320/2284) in the best-performing treatment group (the group that took a medium dose).

Basically, if I was a random member of the study population taking the placebo, I would have had a 16% chance of having another heart attack, stroke or dying in that 4 year period. Crudely, this is a risk of about 4% per annum.

If I’d been taking the study drug, I’d have had a 3.5% risk per annum.

So what does a reduction of 0.5% mean? It means (very roughly) that to prevent one additional heart attack, stroke, or death, we need to treat 200 people for one year.

Now, I have no idea what price this drug will be sold for, but new drugs of this kind often command prices of $10,000 per year, or more.

That’s $2 million dollars.

And there was no reduction in ‘all cause mortality’ in the treatment group.

That means that the ‘halving of cancer deaths’ was balanced out by increases in deaths from other causes, such as serious infections.

That’s $2 million dollars without saving a life.

So, what else can we buy for $2 million? What is the opportunity cost of funding this drug?

Well $2 million is a lot of dietician appointments, a lot of personal trainers, a lot of quit smoking programmes, a lot of health insurance, a lot of income protection insurance, a lot of cardiologists.

Basically that’s a lot of prevention and resilience against future problems that could benefit all 200 of those patients, not just the one who would go on to have the extra heart attack.

At Adapt Research we provide objective health research analysis and plain writing. Send us your question here.

How to survive the next big pandemic

Vaccine2

The New Zealand Ministry of Health has recently published its 2017 pandemic action plan. Flicking through it I noted that it tends to cite other governmental publications rather than academic sources.

Of course, many of the publications cited may well cite the academic literature themselves, but I decided to take a quick independent look at what has been published since the 2009 H1N1 pandemic.

So this afternoon I searched PubMed for ‘pandemic, public health, virus’ and limited the results to the last five years, and review articles only. This turned up 354 results. I read the titles and selected 45 for abstract review. Please note, this is a quick look, so I have not read any full-texts.

The findings of these recent reviews can be collated under the following seven headings:

Travel Restrictions

  • Travel restrictions delay pandemics if implemented within 6 weeks, but only reduce case numbers by 3%.

Although note a modelling study by Adapt Research Ltd, which suggests good cost-benefit for border closure in island nations.

Vaccines

  • Vaccines are effective but there are cognitive facilitators and barriers to vaccination
  • Some cross-protection occurs between strains, so any vaccination might be better than no vaccination
  • There are attempts around the world to develop a universal vaccine effective against all pandemic influenza
  • The Influenza Risk Assessment Tool (IRAT) can be used to prioritize vaccine development for those strains with most pandemic potential

A problem with vaccines is the clear difficulty in changing knowledge, attitudes and behaviours related to influenza and influenza vaccination, particularly on the scope and scale needed to greatly improve uptake.

Modeling the spread

  • Modelling is potentially useful in real time, but its effectiveness still needs evaluating in real pandemic situations
  • During the 2009 influenza pandemic modeling work struck problems with data availability, dissemination, heterogeneity, and unclear assumptions
  • Modelling might be able to be used to identify higher risk populations on whom to target interventions
  • Modeling can provide a quantitative estimate of the impact of various interventions

The challenges that modelling faced in 2009 were: (i) expectations of modelling were not clearly defined; (ii) appropriate real-time data were not readily available; (iii) modelling results were not generated, shared, or disseminated in time; (iv) decision-makers could not always decipher the structure and assumptions of the models; (v) modelling studies varied in intervention representations and reported results; and (vi) modelling studies did not always present the results or outcomes that are useful to decision-makers.

Behaviour

  • Precautionary behaviours are less frequent than expected or intended given the threat during a pandemic
  • Difference in behavior between populations within countries is marked (this suggests targeting interventions might be done better, and a one-size-fits-all response may not be appropriate)
  • Misconceptions about risk are common and vaccination uptake is low
  • Risk communication needs to be tailored to the perceptions/behavior being seen in real time, monitoring social media might help
  • Hand washing has modest efficacy, and dental hygiene may be useful, but other interventions have not been fully assessed
  • Healthcare workers’ willingness to work in a pandemic is variable (this will need to be accounted for in any planning/workforce assumptions)
  • Effectiveness of school closures was unclear in a Japanese review

It seems that a comprehensive, longitudinal study is needed to clarify the effects of school closure and other public distancing measures during pandemics.

Communication

  • Emergency response planners should consider leveraging social media to track population beliefs and behavious in real time, and consider individually tailored engagement and communication
  • There are a number of potential predictors of behavioral compliance with preventive recommendations, these might help focus interventions

Cost-effectiveness

  • Cost-effective are: hospital quarantine, vaccination, antiviral stockpile usage
  • Not cost-effective are: school closures, antiviral treatments, social distancing (at $45,000 willingness to pay per QALY)
  • These interventions are potentially more cost-effective the more severe the pandemic
  • However, cost-effectiveness modeling in the local context is needed.

Ethics

  • Pandemic plans need ethics frameworks that can be used in unique infectious disease pandemic situations, yet most pandemic plans copy and paste ethical approaches from previous influenza plans.

To summarize: How ought we prepare for the inevitable next pandemic?

Before the pandemic hits:

  • Seasonal influenza vaccine
  • Personal protective gear stockpiles
  • Strategic drug stockpiles
  • Risk communication strategy in place
  • Plan for modelling and data needs
  • Regional cooperation plan
  • Plan to research during pandemic (to inform future plans)

During the pandemic

  • Real time PCR for diagnosis (recommended by the CDC)
  • Case surveillance
  • Surge capacity ensured
  • Antiviral drug delivery
  • Risk communication implemented
  • Adherence to strict sanitary and hygienic measures
  • Regional collaboration and cooperation
  • Focus on high-risk groups
  • Data collection and research to inform future response

The above information is consistent with research published in the last five years, and ought to be considered for further research or evaluation, or inclusion in any local, national or international pandemic plan.

Artificial Intelligence: Fake news, digital evolution and free will

puppet_original_9439

The moot at a recent Oxford Union debate reads, “This house believes that fake news is a serious threat to democracy and truth.” the fact is, it’s far worse than that.

Artificial intelligence (AI) is poised to catastrophically transform the information ecosystem and in the process destroy all semblance of truth, fact, knowledge and our ability to act freely and autonomously.

This is because AI will win the fake news game, and it will evolve and adapt to perfectly exploit the psychological weaknesses of human beings. This is foreseeable, and unleashing such intelligences to fight human information wars will be negligent at the least and malicious at worst.

Reasons why AI will destroy the notion of truth

Fake news can already be engineered to discredit journalists and cause real life political demonstrations over issues that do not exist. A Trend Micro report claims it costs $200,000 at present to cause such events.

Autonomous agents now account for 45% of social media posts in some countries. Masses of fake content can give the impression of popularity and cause conformist behavioural effects.

Fake news resembling, at first glance, legitimate sources is now widespread. This can drive belief and behavior through prestige-based psychological effects.

Social media databases harbor vast quantities of psychological and preference information about billions of humans. This data can be exploited to psychologically profile every user on Earth.

Human minds are hackable:

Human psychology is flawed and open to exploitation and manipulation as classic experiments in power and authority, conformity, bias and ideology demonstrate.

AI will learn to harvest this social information, and draw associations and temporal connections between information and behaviour. A recent systematic review details progress to date in using social media to predict the future.

Rival human factions will deploy such AIs to create, distribute, target, and deploy fake content individualized to the susceptibilities of individual human beings on a massive scale. The limits on human productivity will not apply to content generated by AI. The output will be unimaginably vast.

These AIs will be programmed as swarms of intelligences able to evolve and adapt to the defenses used by the AIs of rival human factions. Fake news spam identifiers will struggle to keep up in this evolutionary arms race.

Almost all Internet traffic and content will become AI generated.

Humans will be misled into beliefs and courses of action many steps ahead of being aware they are being manipulated, if even aware at all.

Human beings will fade into the background awash in a polluted information ecosystem, unable to discern fact from fiction or reality from revision.

We will lose all ability to act on information and evidence and thereby lose all freedom and autonomy.

This is the real threat of AI.

Adapt Research promotes the importance of clean information and the notions of risk management, and evidence-informed policy.

To discuss collaborating to understand the issues of the social and institutional threat from AI, fill out our contact form here.