This site uses cookies.

Artificial Intelligence and the Professions 2018 - Dr Mark Burgin

19/12/18. Dr. Mark Burgin BM BCh (oxon) MRCGP explains how the recent advances in AI such as convolutional neural nets will change the everyday work of the professional. (1)

Change has always been a concern for workers from the days of the Luddites through the invention of the printing press and now the computer.

Those who believed that these inventions meant the end of work have been proved as wrong as those predicting free unlimited power from nuclear fusion.

The limiting factor in work is the ability and desire of people to consume what is being produced and that limit appears unlikely to be reached within the medium term future.

Improving AI has meant changes to the way that we work from the email which has speeded up communication of documents to increasing amounts of electronic storage.

The professionals who adapted their practice to incorporate previous technological advances and ignored the hype now have the most successful organisations. (2)

Data management

Transferring the responsibility for data input from the professional to the client using an online form is no more likely to be successful than using a paper questionnaire.

The latest developments in pattern recognition have allowed voice recognition and OCR (optical character recognition) to join the keyboard as methods of data entry.

A new approach where large documents of thousands of pages can be automatically read, and the likely locations of important information can be identified may prove to be useful.

As humans have less than 100% accuracy the machine’s error rate may be acceptable for many purposes allowing the costs to fall with reduced workload.

There will be further developments with the possibility of automatic transcription of conversations between professional and client and document writing systems.

Advertising

Having an online presence allows a professional to scale up their activities and take a larger share of the market thus pushing out those who are not able.

Simply taking the market does not mean that the professional can deliver as for instance, outsourcing with technology-enabled business services has not always been successful.

Advertising can dupe both the professional and the client for a while so that a rapid expansion is often followed by an equally rapid collapse when that service is not delivered.

Shopping has only been able to move from the high-street to online by offering free returns and refunds for items with complex choices such as clothes at huge cost.

Professional services are all complex so there will be many areas where a client-professional interaction will continue to be required to ensure a reasonable outcome.

Professional judgement

Machine pattern recognition using convolutional neural nets is only effective when the input stays the same as they learn from many examples of the same thing.

This type of optimised system is ideal for reading eye scans or typed documents but is less good at tasks such as handwriting where the input is not consistent.

Systems that can take several factors and assign a score have been used in banking and medicine for many years, but the predictive values can be no better than chance.

Chess computers can beat humans but a human who is good at working with computers can beat a human who is good at chess when both are assisted by a computer.

The next generation of professionals will not be replaced by computers but will have new and powerful tools to improve the services that they can offer.

Conclusions

AI does represent a threat to professionals and their job roles as it has already changed what is possible and increasingly what will be expected from professionals.

People will continue to want more, and the professionals are unlikely to keep up with their client’s expectations, even with machine augmentation.

Those who do not learn to work with the new tools may find themselves at a disadvantage in terms of the price that they can offer their services and the speed that they can work.

Neural networks are an important advance in pattern recognition and novel applications of this technology will be developed but there are limits to neural nets.

The professionals work could be transformed by systems that improve client-professional communication but neural nets do not offer any solutions to this type of problem.

Doctor Mark Burgin, BM BCh (oxon) MRCGP is on the General Practitioner Specialist Register.

Dr. Burgin can be contacted This email address is being protected from spambots. You need JavaScript enabled to view it. and 0845 331 3304 website drmarkburgin.co.uk

References

  1. ujjwalkarn 2016 An Intuitive Explanation of Convolutional Neural Networks www.ujjwalkarn.me

  2. Mc Kinsey and Company 2017 A Future That Works: Automation, Employment, And Productivity www.mckinsey.com

    Image ©iStockphoto.com/miriam-doerr