DigitalLawUK Managing Director Peter Wright was invited to attend a Round Table meeting on Artificual Intelligence, Machine Learning and the Professions at the Royal Society on Wednesday 13 July. In advance, participants were able to produce a concise provocation paper for discussion. Please see below the paper that was submitted below looking at the effect on the Legal Profession:
The vast majority of Law firms are resistant to the idea of utilising Artificial Intelligence &
Machine Learning. As a result new entrants to the market are making significant gains, for now in
providing help and advice in areas that Law Firms have stopped providing practical advice on in
recent years due to their perceived lack of profitability following the withdrawal of legal aid.
Chief amongst these is 19 year old Joshua Bowder, who has helped 160,000 people overturn
parking tickets using his DoNotPay chat bot. A 2nd year student at Stanford University, Joshua
consulted machine learning experts to create the bot over a three month period. The App is only
limited in its use by the parking laws that it has assimilated. First set up to cover London,
DoNotPay has also had success in New York and will next roll out in Seattle.
DoNotPay asks the user a series of questions to see if the ticket can be appealed, trying to
identify whether there may be a basis to appeal based on signs not being visible or if the driver
was driving to hospital due to a medical emergency. The system utilises Google Maps to verify
locations where vehicles are parked as many users fail to take photographs at the time that the
ticket is issued.
DoNotPay can also calculate compensation for delayed flights, but Bowder has plans to expand
the range of free advice that the system can provide in order to help vulnerable groups such as
refugees, and will employ IBM's Watson in order to translate.
It must be noted that DoNotPay is currently being run on a not for profit basis, although Bowder
plans to recruit several in house lawyers in order to expand his services and will ultimately
require funding to do this.
The Law Society ran an event at it's headquarters at Chancery Lane during London Technology
Week on Robots & Lawyers. It became apparent from an online survey conducted during the
event and during questions from the floor that such ventures are not being considered by
solicitors or their law firms in the short to medium term future, despite the technology now
existing to allow such systems not only to function but to become appealing to a wide range of
potential users.
One issue raised during the ensuing questions for the panel was whether this technology will
reduce the employment opportunities for young lawyers disproportionately compared to senior
solicitors and Partners in Law Firms. This is indeed one potential outcome of the provision of
fully automated and semi – automated legal services. However it is also possible that by using
systems like DoNotPay, overseen by comparatively junior staff, would actually reduce the need
for supervision and management from more expensive members of staff, which has indeed been
a pattern in many law firms over the last 20 years as increasingly sophisticated case
management systems have allowed armies of comparatively low–skilled paralegals to undertake
bulk – delivered work such as conveyancing and personal injury litigation work that in previous
generations would have been handled by fully qualified and consequently more expensive staff.
The addition of Alternative Business structures to the market may facilitate innovation of the
type typified by DoNotPay with non-lawyer owners of law firms being willing to invest in such
technology where more traditional firms lack the vision and strategic planning to make the use
of such systems profitable and worthwhile in the short to medium term compared organisations
that are run on more entrepreneurial lines.
The first fatality last week sustained by a vehicle while running in “Autopilot” mode raises
significant legal issues. Insurers to date have broadly welcomed the advent of semi automated
and fully automated driver less cars as a way of reducing accidents and premiums. However,
where a fatality is caused by a car as opposed to its driver, depending on the data recorded by
the automated systems at the time of an accident, liability may attach to the vehicle
manufacturer rather than the driver. Liability could also become an issue where a vehicle has
been the victim of a cyber attack and has been “hacked”. Such an attack could lead to different
aspects of the vehicles control systems being compromised. Where such an attack causes an
accident, loss or damage, would the manufacturer be liable for any weaknesses in their system
having been exploited by a attacker in such a manner?
From a user perspective, the advent of semi-automated and fully-automated vehicles has the
potential to massively reduce the number of road traffic accidents. However, will the public be
as tolerant of accidents caused as a result of a automated system compared to a accident
caused by more traditional human error? Such reluctance is likely to prove one of the only
brakes in the adoption of such technology across society as a whole.
The post Artificial Intelligence, Machine Learning and the Law appeared first on DigitalLawUK.