This has reared its ugly head again.

http://politics.guardian.co.uk/homeaffairs/story/0,,2050811,00.html

thanks Paul - i was wondering what lay behind the news management yesterday - west midlands regional news made a big splash about a fraudulent DLA claimant, who received a jail sentence yesterday, and the 'news' of £27m lost in benefit fraud in the west mids was handily equated to the numbers of police on the street...cynic that i am.

Capita will do very nicely out of it, i expect...

also in the news yesterday was Mr. Reid's astonishing idea of linking loudspeakers to CCTV camera spies, who presumably can tell us to pick our feet up and take out hands out of our pockets as we go about our business...

to think orwell fought terminal tb to write his 1984 masterpiece before the reaper's arrival...

this er...new contract with the people...sucks, eh?

You can pick up cheapo versions of this kind of thing on the web (there's even a free one to use with Skype).

Why not announce a pilot where benefit officers calls will be assessed using this to decide whether they really believe what they're saying.

On a more serious note, if you look at the British and European Polygraph Association (BEPA)website at http://www.europeanpolygraph.org/index.htm you will see they emphasise how much training is needed to use these things properly. It might worth getting a statement from them of their views on these pilots to use as evidence where relevant.

"Mr Hutton said: "This technology-based process aims to tackle these fraudsters while speeding up claims and improving customer service for the honest majority.""

So this means if you pass the lie detector test on the phone the DWP will award your benefit without further verification??????? My A***E.

And just think, they could set the system up to detect for 'call centre claim rage' and divert irate claimants and welfs to an answer machine. "I'm sorry, your call can't be taken right now, we have no one in the building with the competence to deal with your claim in an efficient and lawfull manner. Please leave your rant after the tone....Beeep"

I'm looking foward (ehem) to the decision notices of the future. "We have decided that you cannot get the mobility component of Disability Living Allowance. This is because when you told us that you had lost both of your legs and that your career in the royal ballet was in ruins, our computer told us that your voice contained stress inflexions. If you want to appeal please do so in one month, we recomend when you phone that you remain calm and use soothing tones"

........Hmmmmm - I know it's the day before a bank holiday and I'm probably not at full speed - but I cannot for the life of me fathom what lies are supposed to be caught during telephone claims.

Have they never heard of the word "verification". I thought the whole point of verification was to make people prove details of the claim.

Where does fraud from lying come into it if all aspects of the claim are proved?

Most O/Ps & fraud I come across is by claimants failing to notify changes. No amouint of technologegy during phone claims will catch these.

Somebody way better informed than me on the badscience website has posted on this subject on their forum

"Most of those voice stress pattern analysis/lie detection software packages are based on a combination of fuzzy logic and Hidden Markov Models. HMMs alone have a fairly bad accuracy rate; the addition of fuzzy logic can boost it to about a 68% accuracy. However, even the very best are likely to come up with about 30% false positive/false negative (with false positive being far more common); the "98% accuracy" that many of these companies advertise is a huge load of crap, which is gotten by using "tests" in which the spoken words are carefully ideal for detection. In other words, not real world by any stretch of the imagination."

No, I don't know what a Hidden Markov Model is either but it appears that the software will have more false positives than false negatives (ie more people telling the truth will be marked as potential liars than vice versa).

Still it got that nice John Hutton in the news...

David

Good point, David. I see from the Digilog website that although they purport to answer the question " Do you think that VRA has been subjected to enough testing and scientific research?" their account becomes evasive just when the subject of false positives comes up, and they resort to the contention that they must be getting it right as otherwise their customers would be snowed under with complaints.

They also say that they have provided links to supporting research. All I can find is testimonials. Am I missing something?

And would digilog have been able to spot the fib in this news story?

http://www.rightsnet.org.uk/cgi-bin/forwardsql/search.cgi?template2=user_details2.htm&output_number=1&sort=news.submission_date+desc&news.ID=1229151814350

According to Terry Pratchett, fuzzy logic is like woolly thinking, only more so.

I strongly suspect that this is another example of classic Labour political manipulation - ie propose something really outrageous, withdraw it when everyone protests and push through something that's only outrageous rather than really outrageous in its place, which is what they wanted anyway.

However, if this goes through, there are some interesting arguments to be had about whether reliance can be placed on the evidence. Of course, a Tribunal is not bound by formal rules of evidence, but the following cases might be useful:

- Fennell v Jerome Property Maintenance (1986) The Times, November 26th: as a matter of principle evidence produced by the administration of some mechanical, chemical or hypnotic truth test is not admissible in evidence.

- Bernal v The Queen <1997> UKPC 18

http://www.bailii.org/uk/cases/UKPC/1997/18.html

The circumstances in which a polygraph test could be admitted would be "rare" and "the arguments against the admission of such evidence are very formidable" (para 39).

- "we do not think that evidence of the results of polygraph tests would be admitted in England in their present state of development .... As technological and medical techniques advance, there is likely to be greater pressure for evidence to be admitted of statements made under 'truth drugs', such as pentathol; or under hypnosis; but these too are at present insufficiently reliable to justify the displacement of the jury's role as principal assessor of the credibility of witnesses. Such authority as there is supports this conclusion ": Phipson on Evidence (16th edn) para 33-13.


DISCLAIMER: This post is intended as a general contribution to the subject-matter under discussion. It is not intended to be relied upon as legal advice. Any person with a similar or identical problem should seek advice from a welfare rights adviser or a lawyer specialising in welfare rights law.

"However, if this goes through, there are some interesting arguments to be had about whether reliance can be placed on the evidence."

Perhaps I'm being naive, but I don't know whether they are far enough gone to treat this woo-woo as evidence. My impression is that they intend to use it to pick out some claimants for closer scrutiny. The great danger is that it will drive off honest claimants who resent the offensive implication that they are lying.

That's also been my reading of this new initiative. Because the vast majority of interaction between JC+ and claimants will be byway of telephone, the aim is for this system to 'highlight' the people who may be telling porkies, to allow for further investigation of their claims.

Notwithstanding the fact that for both claimants and representatives, dealings on the telephone in connection with benefit claims, as well as issues arising from claims, are mostly a stressful business anyway, you do wonder how this new system will operate for a person with a speech impediment or a difficult accent, for example? And what about someone using Typetalk?

There is an article giving some more background here

There are also notes of speeches from a recent DWP seminar, in partnership with Dods and SPSS, http://www.dwp.gov.uk/welfarereform/seminars/#jan23 about the future of benefit delivery, which has the following from Jim Murphy:

'So my third aspiration, is to see a system where pen cannot be put to paper on a benefit claim, until a work focused interview is completed, and work related activity has already begun. Work search first; benefit second.

At the moment, the application for benefit and starting the process of work related conditionality have to be concurrent for the benefit to be processed within an acceptable time. But in the 21st century, technology should allow us to process benefit claims much faster than we do today. When a bank loan can be applied for in a matter of seconds, when a mortgage can be agreed at the touch of a button on the internet, why does it take 2 weeks to process a benefit claim?

I want to see the time it takes to get benefit halved over the next decade. Through the technological advancements, I am convinced that this is possible. This could then provide us with the opportunity I feel we need to further embed the work first approach into the system. Conditionality requirements could be put in place not just for continued receipt of benefit, but to claim benefit in the first place.

This is not punitive: we would deliver benefits in exactly the same time, or probably a shorter time than we do today. But we would also be giving people the earliest opportunity to focus on their potential and choices within the labour market; and giving the taxpayer a fairer deal for their money.'

So the ambition is to significantly decrease the processing time of benefit claims (a good thing) with the trade-off that benefit claims will not be even accepted until the claimant has proven that they have begun looking for work (not such a good thing for many people, i would posit). And the decision as to whether a person is being honest about their work seeking activity relies on the voice software presumably????

Jim Murphy’s tone is worrying and reflects current trends in welfare provision, which remind me of American style workfare.

There is some very interesting reading at the following link.

http://www-pluto.informatik.uni-oldenburg.de/~also/welarcin.htm

Jim Murphy's tone is worrying that of one who takes for granted the government's divine right to receive national insurance contributions and tax.

i would be surprised if it's the same one he uses when he consults with trade unions leaders. oh what?

somebody please tell me on what basis a person who may have just gone through a fairly traumatic redundancy situation, is to be prevented from making a benefit claim on the national insurance contributions he may well have given to jim murphy and his ilk for a number of decades, before a Jobcentre person gives him permission to? and please tell me how this system might work in the event that we have a recession like the one we had here in the 80's?

and somebody please tell me how not calling it workfare makes it not workfare?

"That's also been my reading of this new initiative. Because the vast majority of interaction between JC+ and claimants will be byway of telephone, the aim is for this system to 'highlight' the people who may be telling porkies, to allow for further investigation of their claims."

If it isn't evidence, it is presumably part of the evidence-gathering process. If it delays the processing of innocent claims ( false positives), there may be an argument for maladministration in the form of unreasonable delay and faulty systems.